The Profitable Bank That Does Not Lend More
This series examines Mauritania’s banking system and a deceptively simple question: why can banks report strong profits without producing equally strong financial intermediation? The focus is not on whether banks should be profitable. It is on the relationship between the earnings of a few dominant banks, the credit they actually extend, and the way bank-specific changes can move aggregate lending. By the end, the reader should be able to distinguish a profitable banking system from a system that reliably broadens access to credit—and judge what the available evidence can support.
The puzzle is not that banks earn money
A bank can make money in several ways without expanding the reach of finance. It can earn a wide spread between lending and deposit rates, collect fees, revalue assets, or charge for taking risks in an economy where borrowers and funding sources are concentrated. Those earnings may strengthen liquidity, provisioning, or capital. That can be valuable: a buffer can help a bank absorb losses and remain open. But a stronger buffer is not the same thing as a new loan to a household, a small firm, or a productive project.
Financial intermediation describes that broader service. Banks collect savings and other funding, assess borrowers, transform maturities and risks, and allocate credit into the economy. Two related questions matter alongside profitability. Access asks who can use formal financial services; depth asks how much credit and other financial activity the system provides relative to the economy. A bank can score well on a narrow efficiency or return measure while the overall system remains shallow or difficult to enter.
The IMF working paper Granular Bank Shocks and Financial Intermediation in a Low-Income Economy uses the Efficiency–Access and Depth Gap, or EADG, to make this mismatch visible. In the paper’s framing, a high gap means that banking-sector profitability or the efficiency dimension is running ahead of access and depth. It is a warning light, not a verdict. A high gap does not prove that market power is the only cause, or that every reported profit is an unearned rent. Wide margins may also compensate banks for operating costs, credit risk, foreign-exchange exposure, or the difficulty of serving a small and concentrated market. The important point is more precise: profit is an outcome to explain, not a substitute for measuring the services the system delivers.
This is more than a distinction between two labels. It is a distinction between a stock of protection and a flow of service. A bank can retain earnings and become better able to absorb a loss without increasing the number of viable borrowers it serves. It can also increase lending for a particular project while current profit barely changes, because the revenue arrives later, the loan carries provisioning costs, or the bank is accepting a different risk profile. Looking only at the income statement would combine these cases and make them appear equivalent. Looking at lending, access, and depth alongside earnings asks whether financial capacity is reaching the economy in a usable form.
The same caution applies to the word “efficiency.” In a financial-development index, a strong efficiency score can reflect high returns and wide margins; it need not mean that customers receive cheap credit or that the system reaches a broad population at low cost. Nor does a low access or depth measure prove that banks are deliberately withholding credit. It may reflect borrower risk, weak information, limited collateral, costly geography, or macroeconomic uncertainty. The value of the EADG is that it identifies a relationship worth investigating. It does not remove the need to investigate the relationship.
That distinction changes how the central question should be phrased. Instead of asking whether profitable banks are “good” or “bad,” ask which balance-sheet movement is connected to new credit. Is aggregate lending responding to changes in the volume and timing of lending decisions? Or is it responding to earnings that remain inside the bank as protection against uncertainty? Those are different channels with different implications for borrowers and policymakers.
Why Mauritania makes the question urgent
Mauritania is a useful case because the paper describes a low-income banking system in which credit is highly concentrated. The five largest banks account for an overwhelming share of the lending dynamics shown in the study, while the remaining institutions play a much smaller role in system-wide credit. That structure creates an exposure that a national average can conceal. If one of the smaller banks changes its lending, other institutions may be able to offset the movement. If one of a few dominant lenders changes its lending, the aggregate system may move with it.
The surrounding balance-sheet conditions make the distinction between earnings and intermediation more consequential. In the paper’s descriptive data, average loan rates are about 11 percent and average deposit rates about 6 percent, implying a lending–deposit spread near 600 basis points. Such a spread can reflect limited competition, risk, and operating costs as well as pricing power; the data do not, by themselves, settle which explanation dominates. The paper also reports meaningful concentration among the largest depositors and borrowers. A bank may therefore be profitable while still managing a narrow set of funding relationships, a concentrated loan book, and risks that make immediate expansion unattractive.
The size distribution reinforces that possibility. Across the bank-month observations, median total assets are about 6.55 billion Mauritanian ouguiya, compared with a mean of 8.76 billion and a maximum above 33.6 billion. Median loans are about 4.20 billion, against a mean of 4.85 billion. Those differences are not proof of a particular competitive failure, but they show why an average bank can be a poor guide to the system. The paper also reports that the five largest depositors account for about 15 percent of deposits on average, reaching 50 percent in some bank-months; the top five borrowers account for a median 21 percent of a bank’s loan book, with the upper quartile at 31 percent. Funding and lending can therefore be exposed to a small number of relationships at the same time that the system is exposed to a small number of banks.
This matters beyond a banking scorecard. In a low-income economy, shallow intermediation can leave small and medium-sized enterprises, rural borrowers, and other underserved groups outside the most useful parts of the financial system. It can also make growth more dependent on decisions made inside a small number of institutions. When credit is concentrated, a local change in lending standards, the approval of a large project, a reassessment of a borrower, or a response to a depositor’s withdrawal can have consequences far beyond that bank’s own income statement.
Concentration is not automatically a crisis, and the paper does not claim that every large bank is unstable. It changes the transmission of shocks. It means that the resilience, governance, and credit standards of dominant banks deserve attention not only because of their own balance sheets, but because they sit close to the system’s main credit channel. It also changes the meaning of a national average. An apparently healthy aggregate profit figure can coexist with a weak flow of new credit if the institutions generating the returns are protecting concentrated balance sheets rather than competing for a wider set of borrowers. Conversely, a short-lived lending expansion at one dominant bank may lift aggregate credit without telling us that access has become durable or broad.
From a country-level gap to a bank-level test
The paper turns this structural puzzle into a test about shocks. A bank’s observed change in lending can reflect common conditions—monetary policy, inflation, broad money, or exchange-rate movements—as well as characteristics specific to that bank. The study asks what remains after those influences are taken into account. That remainder is treated as a bank-level, idiosyncratic innovation. When the remainder comes from a large bank, it receives more weight in the aggregate measure than the same-sized remainder from a small bank.
The comparison between lending and profitability is the key. A lending shock is an unexpected change in the pace of credit expansion after the study’s controls are applied. A profitability shock is an unexpected change in earnings after the corresponding controls are applied. The two can occur together, but they need not. A bank can experience an earnings increase and retain it as liquidity, provisions, or capital. It can also expand lending because of a large project or a change in internal risk limits without producing a comparable rise in current profit.
To examine that difference, the authors use confidential monthly supervisory data covering the post-COVID period from June 2020 through June 2024. The final panel contains 15 banks and 49 monthly observations, with two banks excluded for the reasons described in the paper; the aggregate regressions use 48 observations. The study then relates size-weighted bank-level residuals to aggregate credit growth and compares them with an equal-weight average. This design does not prove every cause of profitability, and it is not a complete causal-instrument strategy. It asks a narrower but important question: after common conditions and observed bank characteristics are removed, do the remaining movements at dominant banks help explain short-run changes in system credit?
That narrower question supplies a useful reading rule for the rest of the evidence. When profits rise, one should ask whether loan volume rises, who receives the additional credit, and whether the change is large enough to alter system depth rather than merely one bank’s balance sheet. One should also ask whether the time horizon is long enough for earnings to become lending, whether the bank is using income to provision for existing risks, and whether smaller institutions can offset or amplify the dominant bank’s movement. None of these questions is answered by a return measure alone.
The paper’s design is valuable precisely because it does not pretend to answer everything at once. It separates the bank-level movement from the common macroeconomic background and then tests whether the size-weighted remainder tracks aggregate credit. That makes the result relevant to supervision and policy, but it also sets a boundary: the study can show where short-run variation is concentrated more clearly than it can identify the full structural reasons why Mauritanian banks earn what they do.
That is the question this series follows. The answer depends on keeping three ideas separate: what banks earn, what they lend, and how concentrated the system is. Once those ideas are separated, high profits stop looking like a complete measure of financial development. They become one piece of evidence whose meaning depends on who receives credit, how much is supplied, and whether the institutions with the greatest influence are actually moving funds into the wider economy.