Nepal’s banking sector has deepened rapidly, yet credit has concentrated in collateral-backed, often non-productive uses while excess liquidity and subdued productive-sector demand persist. Examined strictly from a banking-law standpoint, this disconnect is not a gap the law fails to address but a predictable product of the incentive structure the law deliberately creates. The framework – the Nepal Rastra Bank Act, 2058, the Bank and Financial Institution Act, 2073 (BAFIA), Nepal Rastra Bank’s Unified Directives, the Bank and Financial Institution Debt Recovery Act, 2058 with its 2059 Rules, the Secured Transactions Act, 2063, and the Banking Offence and Punishment Act, 2064 – is doctrinally coherent and prudentially sound. But it operates in two layers that pull against each other. An underlying incentive system – risk-weighting, liquidity design, collateral realities, provisioning triggers, and recovery frictions – systematically favours secured, government, and low-risk-weight exposures. A pro-productive apparatus of sectoral quotas, interest caps, preferential provisioning, flexible collateral, raised single-obligor bands, and refinancing sits on top of it as a set of overrides. The mandates compel a floor of productive credit; they do not reorient the marginal incentives governing lending above that floor. This article traces that tension through each layer of the framework, and argues that it is the faithful working-out of a statute which, by design, subordinates credit allocation to financial stability.
The Statutory Frame: Stability First, Development Through It
The architecture begins with a deliberate ordering of objectives. Section 4(1) of the Nepal Rastra Bank Act, 2058 fixes the central bank’s aims as monetary and price stability (s.4(1)(a)), stability of and public confidence in the financial sector (s.4(1)(b)), and a secure, efficient payment system (s.4(1)(c)); economic development enters not as a co-equal aim but as a consequence, with s.4(2) directing the Bank to support Government economic policy only where doing so does not adversely affect those primary objectives. BAFIA’s Preamble echoes the hierarchy. The design thus subordinates credit allocation to prudential soundness as a matter of law – the first structural reason the framework leans toward safety: where the two conflict, the statute privileges stability.
On that foundation the Act builds NRB’s regulatory reach. Section 5(1)(ch) empowers it to license and supervise BFIs; s.79(1) grants “full authority” over their functions, and s.79(2) makes compliance with its directives a legal “duty” of every institution. BAFIA s.131 supplies a concomitant directive-making power. The Unified Directives – the 2081 compilation for Classes A, B, and C and the 2082 compilations for Class D microfinance and for the Infrastructure Development Bank (NIFRA) – are issued under s.79 and carry the force of law, enforceable through BAFIA s.99 (licence suspension or cancellation) and NRB Act s.100 (personal, non-reimbursable fines on directors and officers). The practical consequence is that the incentives governing credit allocation are set primarily in binding sub-legislation rather than statute, and can be adjusted administratively.
Classification, Permissible Activities, and the Architecture of Intermediation
BAFIA s.37 classifies licensed institutions into Classes A (commercial banks), B (development banks), C (finance companies), and D (microfinance), with NIFRA outside the letter scheme. Section 49 tiers permissible activity: Class A enjoys the broadest authority, including project, hypothecation, and consortium financing and foreign exchange; Classes B and C are narrower, several functions requiring prior NRB approval; Class D is confined to group-based micro-credit; and NIFRA under s.49(5) is restricted to infrastructure-project lending and equity, barred from retail or consumer banking. Section 50 draws the outer boundary: no trading in goods except to realize security, no lending against a BFI’s own shares, no credit to directors, 1%-or-greater shareholders, executives, or their families (the connected-lending bar), and no breach of the single-obligor limits NRB prescribes. The differentiated capital and reserve obligations that follow from classification determine which institutions can hold which risks.
| Requirement | Class A | Class B (nat’l) | Class C (nat’l) | Class D (MFI) | NIFRA |
| Min. paid-up capital | Rs 8 bn | Rs 2.5 bn | Rs 800 mn | Rs 100 mn | Rs 20 bn |
| Capital fund ratio | 11% | 11% | 10% | 8% | 11% |
| Cash Reserve Ratio | 4% | 4% | 4% | 2% / 0.5% | 1% |
| Statutory Liquidity Ratio | 12% | 10% | 10% | 4% | 12% |
Capital figures per Unified Directive 2081, Directive 21, Clause 16(1) (Classes A/B/C) and Unified Directive 2082 (Class D and NIFRA); capital-fund ratios under CAF 2015 (A/B/D/NIFRA) and CAF 2007 (C). CRR and SLR per Unified Directive 2081, Directive 13, and the corresponding Class D and NIFRA Directive 12. Class D wholesale MFIs face a Rs 600 million paid-up minimum.
The Credit-Approval Discipline
At the transaction level the law fixes a repayment-and-security test that structurally privileges collateral. BAFIA s.55(1) permits credit only after the BFI satisfies itself as to repayment capacity, purpose, and adequacy of security, and s.56 requires a monitoring schedule to verify utilization. The Directives operationalize this: Directive 2, Clause 2 permits disbursement or renewal only after analysis of repayment capacity, future cash flow, and source of income; Clause 1(gh) requires that approval to any firm or company rest on financial statements held in the Inland Revenue Department’s Integrated Tax System; and Directive 5, Clause 4(6) requires Class A and national-level Class B institutions to maintain an internal Credit Risk Rating System, with mandatory external rating for borrowers using Rs 500 million or more (Directive 2, Clause 34). Two features narrow the productive funnel: the Integrated Tax System requirement effectively excludes informal or under-documented MSMEs that cannot present tax-verified financials, and the security-adequacy test anchors lending culture to realizable collateral rather than to project cash flows.
Concentration Limits and the Single-Obligor Regime
Concentration control is where the framework most visibly tries to steer credit toward productive uses – and where its penalty structure creates the sharpest disincentive. Directive 3, Clause 2 sets a general Single Obligor Limit (SOL) of 25% of primary capital for fund- and non-fund-based exposure to a single customer or related group, raised to 30% for specified productive sectors (exports, SMEs, transport infrastructure, pharmaceutical manufacturing, agriculture, tourism, and domestic-input cement and iron) and to 50% for hydropower, transmission, renewable energy, and cable-car projects – an explicit accommodation for large productive exposures. Margin-nature share-pledge lending is separately capped at Rs 25 crore per borrower. The related-group definition is broad: 25%-or-greater cross-shareholding, family and ownership links, and cross-guarantees all consolidate borrowers into one group, though majority-government entities are treated separately.
The enforcement is where the disincentive bites. Under Directive 3, Clause 10, any exposure exceeding the SOL requires a 100% loss provision on the excess, and the Capital Adequacy Framework 2015, s.6.4 adds 10% of the excess exposure to risk-weighted exposure. Sectoral concentration is capped at 40% of the total portfolio, with board ratification required where a single sector exceeds 100% of Tier 1 capital. So even a well-secured, economically sound large project – precisely what the raised 30–50% bands are meant to enable – becomes prohibitively expensive the moment it nudges past the limit, since the penalty is a full provision plus a capital surcharge regardless of collateral quality. The framework simultaneously invites large productive exposures and penalizes them at the margin.
Capital Adequacy: The Risk-Weight Incentive That Shapes Allocation
If any single mechanism explains the allocative tilt, it is risk-weighting. Because risk-weighted exposure determines how much capital a bank must hold against a loan, and capital is scarce and costly, the risk weight attached to each asset class is in effect a price signal set by the regulator. The schedule points one way. Claims on the Government of Nepal and NRB carry a 0% weight, so a bank may hold unlimited government securities without consuming regulatory capital. SME loans in agriculture, IT, or manufacturing up to Rs 30 million carry a preferential 60% weight where fully secured with a 25% margin or DCGF-guaranteed (CAF 2015, Annex 1.1, Clause 3.3(j)(1)); residential housing 60%, staff housing 50%. Against these, unrated domestic corporate claims carry 100%, and the sectors most associated with productive risk-taking – venture capital, private equity, unlisted equity, and personal overdrafts – carry 150%.
| Exposure class | Risk weight | Capital-allocation effect |
| Government of Nepal / NRB securities | 0% | No capital consumed; unlimited holding |
| SME (agri / IT / mfg, secured, ≤ Rs 30 mn) | 60% | Preferential; encourages small productive lending |
| Residential housing | 60% | Favoured over general corporate |
| Unrated domestic corporate | 100% | Baseline productive-lending cost |
| Venture capital / private equity / unlisted equity | 150% | Penalized; discourages growth-risk finance |
| Exposure exceeding Single Obligor Limit | +10% of excess to RWE | Capital surcharge on large exposures |
The gradient runs away from exactly the higher-risk productive lending a developing economy most needs and toward government paper and secured exposures. NRB partly counteracts this through Credit Risk Mitigation relief (0% weighting for GoN- or multilateral-guaranteed exposures) and the raised SOL bands, but the baseline remains oriented toward capital preservation – which is precisely what the s.4 stability mandate contemplates.
Liquidity Design and the Pull Toward Government Paper
The liquidity framework compounds the tilt. NRB Act s.5(1)(i) and BAFIA s.46 authorize the reserve and liquid-asset mandates that Directive 13 specifies: a 4% Cash Reserve Ratio and, for Class A, a 12% Statutory Liquidity Ratio (10% for Classes B and C), with government securities the primary eligible SLR asset (Directive 13, Clause 3). The effect is that a large, legally-mandated share of every balance sheet must sit in instruments that are simultaneously SLR-eligible and 0%-risk-weighted – a double incentive to hold government debt that private-sector lending cannot match, since private loans neither satisfy the SLR nor enjoy the capital efficiency. NRB imposes a countervailing discipline through the 90% Credit-to-Deposit ceiling (Directive 5, Clause 6(6)), penalized daily at the bank rate, together with structural-liquidity gap analysis across five maturity buckets, a mandatory ALCO, and a contingency funding plan. But within those limits government paper is the path of least regulatory resistance – the mechanism most directly implicated in persistent excess liquidity: when productive demand is weak, the framework offers a capital-efficient, SLR-satisfying, penalty-free place to park deposits rather than pushing them toward productive credit.
The Mandate Side: Priority-Sector Quotas and Productive Incentives
Against these market-following incentives the Directives layer an explicit command-and-incentive regime. Directive 15 sets graduated targets: Class A must reach 15% of total credit in agriculture, 10% in energy, and 15% in SMEs by end-Ashad 2084; Classes B and C must reach 20% and 15% in prescribed sectors; and Class D must direct one-third of total credit to agriculture (Directive 16, Clause 3(a) of the 2082 MFI compilation). Shortfalls are fined at the shortfall multiplied by the prevailing bank rate, converting the target into a cost-avoidance incentive. Alongside the quotas the framework offers genuine carrots:
• Interest-premium caps: productive loans up to Rs 2 crore in food production, animal husbandry, fisheries, IT, and domestic manufacturing at no more than base rate plus 2% (Directive 15, Clause 39(2)); electricity-export projects at base rate plus 1% for five years (Clause 39(3)).
• Preferential provisioning for agriculture: long-term commercial fruit and cash-crop lending at reduced pass-loan provisioning of 0.2% / 0.6% / 1.10% across years one to three (Directive 2, Clause 7), against the standard 1.10%.
• Flexible collateral: agricultural land without motor-road access accepted for commercial-farming loans up to Rs 20 lakh (Directive 15, Clause 39(1)); the project and its own assets as primary security for productive projects up to Rs 5 crore (Directive 2, Clause 10(11)(cha)).
• Refinancing: NRB Act ss.49–50 and the Refinance Procedure, 2077 (Directive 21, Clause 27(a)) provide lower-cost funds for on-lending to productive sectors.
• Raised SOL bands and CRM relief: the 30–50% single-obligor bands and 0% risk-weighting for guaranteed exposures enable large infrastructure and energy financing.
This is a substantial apparatus, and it shows the framework is not indifferent to allocation. Its limitation is structural, not textual: the mandates are overrides bolted onto an incentive system that points the other way, so banks meet quantitative targets to avoid penalties while the marginal, capital-and-provision-driven logic of the rest of the framework keeps rewarding secured, low-risk-weight, government-adjacent lending. The mandates compel a floor of productive credit; they do not shift qualitative appetite above it.
Collateral, Recovery, and the Persistence of Immovable Security
The collateral regime is doctrinally modern but operationally weighted toward land. The Secured Transactions Act, 2063 establishes a genuinely broad movable-collateral framework: s.2(jha1) defines movable property to include future property, accounts receivable, inventory, bank accounts, and intellectual property, enabling floating charges and receivables and inventory financing that older law could not support. Perfection runs by registration of a notice at the Registry (s.4) or by possession or control (s.11(8)–(9)); priority by time of perfection (s.12); a perfected interest outranks unsecured claims including tax and wage claims (s.29ka); and enforcement is expedited through a self-help remedy allowing repossession of movable collateral without judicial process on default (s.48(1)). On paper this is exactly the infrastructure needed to lend against business assets rather than land.
The surrounding incentives pull the other way, and recovery law reinforces them. General real-estate LTV is capped at 50% (Directive 3, Clause 11(4)(ka)), but immovable property remains the security that satisfies the s.55(1) adequacy test most readily, while collateral-insufficiency triggers immediate loss classification and 100% provisioning (Directive 2, Clause 3(ja)); movable-asset lending, by contrast, carries valuation, monitoring, and enforcement uncertainties the risk-weight and provisioning schedules do not reward, and s.13(1) of the Banking Offence and Punishment Act criminalizes over-, under-, or wrong valuation, sharpening caution around harder-to-value assets. On the recovery side, BAFIA s.57 lets a bank auction pledged security on breach or diversion of funds and take over an unsold asset; the Debt Recovery Act, 2058 supplies a specialized Debt Recovery Tribunal (s.15) that can freeze assets and, through the Debt Recovery Officer, seize and auction the borrower’s or guarantor’s other property (s.25), with the 2059 Rules requiring three auction attempts before compelled takeover (Rule 29). But takeover converts unsold collateral into Non-Banking Assets, which Directive 8 requires be provisioned at 100% if not sold within the prescribed time and which reduce capital through deduction from Tier 1 – so enforced-but-unsold security becomes a capital drain, not a recovery; and the multi-attempt auction and Tribunal adjudication introduce delay during which the exposure sits classified and provisioned. The rational response, visible in the framework’s own incentives, is to lend against liquid, easily-realizable immovable collateral in prime locations and to avoid exposures – many of them productive – whose security is specialized, movable, or slow to sell. The Secured Transactions Act’s modern provisions thus remain under-leveraged because the surrounding prudential incentives do not reward it.
Classification, Provisioning, and Distressed-but-Productive Credit
Loan classification translates credit quality into immediate profit-and-loss consequences, and its design discourages continued support of productive projects that hit turbulence. Directive 2 grades loans as Pass, Watchlist (one to three months overdue, or failing DTI or debt-to-equity thresholds), Sub-standard (three to six months), Doubtful (six months to one year), and Loss (over one year), with provisioning rising from 1.10% for Pass through 5% at Watchlist to 100% at Loss. The Watchlist trigger is the consequential one: a project facing cost or time overrun, a cash-flow mismatch, or a debt-to-equity ratio worse than 80:20 must be moved to Watchlist and provisioned at 5% (Directive 2, Clause 1(2)) even where the underlying project remains viable – exactly the circumstances that characterize infrastructure and manufacturing during construction and ramp-up. Normal project turbulence thus forces a five-fold jump in provisioning, discouraging origination of long-gestation productive loans.
The response is a double-edged set of restructuring controls. Rescheduling and restructuring provisions allow relief for genuinely distressed borrowers, but NRB has progressively tightened the rules on interest capitalization and evergreening to stop banks masking non-performance – a genuine driver of understated NPLs, and a control that properly serves the stability mandate. The problem is that: tighter evergreening protects asset-quality integrity but reduces the flexibility to carry productive borrowers through temporary distress, reinforcing a preference for shorter-tenor, self-liquidating, well-secured exposures over patient capital. And where a loan is proven to be diverted from its sanctioned purpose, Directive 2 requires immediate Loss classification and 100% provisioning regardless of repayment status – a powerful integrity mechanism that also raises the stakes of any productive lending whose end-use is hard to monitor.
Criminal Liability, Governance, and Lending Capacity
The criminal and governance layer protects allocative integrity but adds to the caution that pervades lending. The Banking Offence and Punishment Act, 2064 criminalizes obtaining credit by false statement (s.7), misuse of a loan for an unsanctioned purpose (s.8), forgery and deception (ss.12 and 12ka), and improper collateral valuation (s.13), with penalties graduated by amount, restitution plus equivalent fines, and an additional year’s imprisonment where the offender is a chairman, director, or CEO (s.15(8)). BAFIA reinforces this through fit-and-proper criteria for directors and senior management, the connected-lending bar (s.50), and mandatory internal-control, audit, compliance, and risk-management structures; the Directives add an ICAAP and stress testing. These controls are individually justified and collectively essential to the stability the statute prioritizes, but their cumulative effect is to raise the personal and institutional cost of lending that is hard to document, secure, or monitor – which describes a meaningful share of productive-sector credit. The criminal exposure of directors and officers for misuse and mis-valuation, in particular, rationally pushes decision-makers toward the safest, most collateralized exposures available.
The same logic closes the loop at the level of lending capacity. BAFIA s.47(2) bars any dividend until accumulated losses are covered, statutory reserves under s.44 are funded, and NRB’s capital and provisioning requirements are met, and NRB’s dividend directives further condition distribution on capital-conservation-buffer compliance. This preserves the capital base on which lending ultimately rests – a pro-stability, pro-capacity design – but it also means that a system carrying rising non-performing loans and provisioning burdens, as Nepal’s has through the recent tightening cycle, sees its distributable-and-lendable headroom compress precisely when productive demand would most benefit from expanded capacity, reinforcing the conservative posture the rest of the framework encourages.
References
- Nepal Rastra Bank Act, 2058 (2002), ss.4, 5, 46 (via BAFIA), 49–50, 79, 100, 110.
- Bank and Financial Institution Act, 2073 (2017) (BAFIA), ss.37, 44, 46, 47, 49, 50, 55, 56, 57, 99, 131.
- Nepal Rastra Bank, Unified Directives, 2081 (Classes A, B, C): Directive 2 (credit, classification, provisioning), Directive 3 (single obligor / concentration), Directive 5 (credit-risk rating, liquidity, CD ratio), Directive 8 (non-banking assets), Directive 13 (CRR/SLR), Directive 15 (priority / productive sector), Directive 21 (capital, refinance).
- Nepal Rastra Bank, Unified Directives, 2082 (Class D microfinance): Directive 1 (capital), Directive 12 (CRR/SLR), Directive 16 (agriculture credit).
- Nepal Rastra Bank, Unified Directives, 2082 (Infrastructure Development Bank / NIFRA): Directive 1 (capital), Directive 3 (single obligor), Directive 12 (CRR/SLR).
- Nepal Rastra Bank, Capital Adequacy Framework, 2015 (updated 2082), ss.3.3, 6.4; Annex 1.1; Capital Adequacy Framework, 2007 (Classes B/C).
- Nepal Rastra Bank, Refinance Procedure, 2077.
- Bank and Financial Institution Debt Recovery Act, 2058 (2002), ss.15, 25; Debt Recovery Rules, 2059 (2002), Rules 15, 29.
- Secured Transactions Act, 2063 (2006), ss.2, 4, 11, 12, 29ka, 48.
- Banking Offence and Punishment Act, 2064 (2008), ss.7, 8, 12, 12ka, 13, 15.