Ghana’s Digital ID Push Reveals Why Faster Banking Does Not Guarantee Wider Credit Access

Ghana's push to modernize financial transactions through a national digital identity system has coincided with faster loan processing, higher approval rates and increased lending activity. Yet the transformation raises a difficult question for policymakers: can banks become more efficient at identifying customers without necessarily becoming more willing to extend credit?

A study titled "Digital Identity and Formal Credit Access in a Developing Economy: Evidence from Ghana's Banking Sector," published in the journal Economies, investigates this distinction. Authored by Emmanuel Awortwe, Michael Kodom, John Kofie Anoku and Akwasi Boahene, the research examines whether Ghana's digital identification reforms changed lending decisions or merely coincided with broader improvements in banking performance.

Drawing on evidence from 23 universal banks between 2017 and 2024, the researchers identify substantial improvements following the Ghana Card mandate. However, their analysis finds no statistically significant association between the timing of individual banks' integration agreements and changes in loan approvals, processing times, repayment or non-performing loans.

The findings expose a fundamental distinction in digital financial inclusion: establishing who a borrower is does not automatically establish whether that borrower can repay.

Banking Became Faster, but the Source of Improvement Remains Uncertain

Ghana's banking reforms sought to address persistent weaknesses in customer identification and financial verification. Previously, banks relied on fragmented documentation, including passports, driving licences and voter identification cards, creating difficulties in verifying identities consistently across institutions.

The Bank of Ghana announced in January 2022 that the Ghana Card would become the sole accepted identification document for financial transactions from July 1 that year. The reform established a common verification framework intended to improve identification reliability and strengthen financial-sector compliance.

The subsequent changes in banking performance were substantial. Average loan approval rates increased from 58.04% during 2017–2021 to 67.99% during 2023–2024, while loan applications rose by 21.7%. Processing time declined by 39.5%, falling from 16.46 days to 9.96 days.

Domestic and foreign-owned banks experienced broadly comparable improvements. Processing times declined by 43.5% among domestic banks and 35% among foreign banks, while approval rates increased across both ownership categories. These patterns indicate that operational improvements extended throughout the banking sector rather than remaining concentrated among particular institutions.

However, improved performance following a policy intervention does not establish that the intervention produced the improvement. Ghana implemented its identification mandate amid exceptional economic instability, including inflation, currency depreciation and domestic debt restructuring, all of which affected the financial environment.

The researchers therefore distinguished nationwide changes from differences associated with individual banks' verification-service agreements. Their models accounted for characteristics specific to banks and economic conditions shared across institutions, allowing closer examination of integration timing.

The resulting estimates showed no statistically significant relationship between documented integration timing and the four principal lending outcomes. Banks executing verification agreements later also recorded improvements, weakening the argument that their individual agreement dates explain the sector's transformation.

Digital Identity Removes Verification Barriers, Not the Uncertainty Behind Lending

The distinction between identification and creditworthiness provides the strongest explanation for why digital infrastructure may not independently alter credit allocation. Identity verification establishes a borrower's authenticity, while lending decisions require additional information about income, existing obligations and repayment capacity.

For borrowers operating in informal economic activities, these requirements remain particularly challenging. Reliable identification may simplify an application, but it cannot independently establish business cash flow, income stability or the financial history needed to evaluate lending risk.

The study consequently distinguishes three functions of financial information: screening customers, allocating credit and monitoring repayment. Improvements in one function do not necessarily generate corresponding changes in the others.

This separation is reflected in the statistical results. Although loan approvals increased substantially across the mandate period, the estimated integration-timing association with approval rates was only minus 0.22 percentage points and was statistically insignificant. Similar results emerged for processing time and portfolio quality.

Additional evidence comes from banks that executed agreements outside the main 2022 adoption period. One institution executed its agreement in December 2023 but had already reduced processing time from 17 days in 2021 to 10 days in 2022 and nine days in 2023.

Another bank executed its agreement in April 2019 but experienced meaningful processing improvements only later, alongside the wider banking sector. These trajectories suggest that recorded agreement dates do not adequately explain when performance improvements occurred.

Nevertheless, the researchers caution that executing an agreement may differ from beginning operational verification. Without information on the first actual verification transactions, the analysis cannot establish whether the absence of measurable effects reflects limited integration benefits or inaccurate measurement of implementation timing.

A further constraint concerns credit-information coverage. The study reports that Ghana Card numbers covered approximately 35% of active loan customers in the database of Ghana's largest credit bureau in December 2023, illustrating the incomplete connection between identification records and borrower credit histories.

The evidence supports a narrower conclusion than either technological success or failure: the available data cannot establish that individual banks' documented integration timing produced the observed lending improvements.

Repayment Patterns Reveal a Possible Benefit That Approval Figures Conceal

Although the study identifies no statistically significant integration-timing effects in its principal models, repayment performance presents a potentially important distinction between granting credit and monitoring borrowers after disbursement.

Persistent identification could strengthen accountability by making it harder for borrowers to escape negative repayment histories through alternative identities. Such traceability may influence repayment incentives even when identification alone does not provide enough information to change initial lending decisions.

A comparison between 20 banks executing agreements during 2022 and two later-integrating institutions found a 5.84-percentage-point difference in changes to on-time repayment. Repayment performance held up better among the earlier integration cohort than among the two comparison banks.

However, the researchers explicitly classify this finding as suggestive. It depends on a comparison group of only two institutions and loses statistical significance after adjustment for multiple outcomes. The evidence therefore cannot establish that digital identification improved repayment discipline.

Portfolio-quality results provide another reason for caution. The study finds no statistically significant association between integration timing and non-performing loan ratios, while Ghana's difficult macroeconomic conditions complicated the assessment of borrower repayment performance.

Lending volumes also require careful interpretation. Average nominal loan size increased by 84.7% between the pre-mandate and post-mandate periods, but the increase occurred alongside considerable inflation and currency depreciation. Higher loan values therefore cannot be equated with equivalent growth in real credit.

Similarly, female borrowers accounted for an average of 38.35% of loans across the observation period, with little difference between domestic and foreign-owned institutions. Aggregate banking information cannot determine whether identification reform improved access among previously excluded women or other underserved borrowers.

These limitations expose an important weakness in evaluating financial inclusion through institutional performance alone. Faster processing, higher approval rates and larger nominal loans do not establish who obtained additional credit or whether lending opportunities became more equitable.

The Next Test Is Whether Verified Identity Can Become Usable Credit Information

For Ghana's financial authorities, the research shifts attention from identity adoption toward the information systems required to make identification useful in credit assessment. Linking verified identities with credit histories, credible income information and repayment records could address constraints that authentication alone cannot resolve.

Such integration could be particularly relevant for informal borrowers whose financial activity is not adequately captured by conventional documentation. However, the researchers do not establish that these complementary measures would necessarily increase approvals; their effectiveness remains an issue for further investigation.

The study also identifies an evaluation challenge. Twenty of the 23 banks executed verification-service agreements during 2022, leaving limited variation in adoption timing. Sensitivity tests produced estimates that changed in magnitude and direction, reinforcing the uncertainty surrounding precise effects.

Future investigation would benefit from monthly banking observations, actual verification-use records and individual borrower information. Such evidence could distinguish administrative integration from operational activity while revealing whether improvements reached previously excluded applicants.

Comparative analysis across African economies could further establish whether the separation between nationwide digital identity reforms and bank-level integration outcomes reflects Ghana's particular circumstances or a broader feature of financial-sector modernization.

For policymakers investing in digital public infrastructure, the findings offer a disciplined framework for measuring progress. Identity coverage, verification efficiency, loan approval, borrower inclusion and repayment performance represent connected but distinct outcomes requiring separate assessment.

Ghana's experience ultimately challenges the assumption that removing a documentation obstacle automatically removes a lending obstacle. National identification infrastructure can establish a reliable foundation for financial transactions, but banks still need credible information about borrowers' capacity to repay.