Kazakhstan's Non-Bank Lending and Private Credit: A Comprehensive Analysis

Kazakhstan's Private Credit Landscape

Kazakhstan is the most deliberately engineered consumer credit market in this series. Not the one with the most rules, although it may be that too, but the one where the regulator has most precisely decided what a lender's balance sheet is allowed to look like.

The licensed microfinance sector holds KZT 3.1 trillion of loans (about US$6.9 billion). Roughly half sits with 214 microfinance organisations; the rest with member-owned agricultural credit partnerships and pawnshops. Every unsecured loan, bank or non-bank, is capped at a 46% all-in annual effective rate. And since 1 January 2026 a microfinance organisation only meets its capital adequacy ratio if loans more than 90 days overdue, including loans already written off, stay below 10% of the portfolio. Lenders have been barred from selling bad consumer debt to collectors since July 2024, and a borrower who is more than a day late on a microloan cannot take new unsecured credit anywhere.

Bad debt, in other words, cannot leave the building. That single design choice shapes reported asset quality, lender behaviour and how a private credit facility to a Kazakh lender has to be structured.

Most readers get two other things wrong. The first is the headline that Kazakhstan's MFO portfolio fell 4% in 2025. It did not shrink: KMF, the largest MFO, became a bank and took KZT 306 billion of loans out of the statistics in a single month, BNK Finance converted too, and the remaining sector grew by about 20%. The second is the headroom argument. Private credit is around a quarter of GDP, which looks under-penetrated. Households are not: debt reached 51% of wages at end-2025, above the level recorded before the 2008–09 banking crisis.

Role of Non-Bank Financial Institutions (ООМФД)

Kazakh law groups three types of lender as organisations engaged in microfinance activity (ООМФД). All are licensed and supervised by the Agency for Regulation and Development of the Financial Market (ARDFM):

• Microfinance organisations (MFOs, mikrofinansovye organizatsii) — 214 licensed in September 2026: eight specialise in car finance, 39 in short-term payday-style loans (PDL) and the rest in "classic" microcredit to individuals and small businesses. This is the investable universe.
• Credit partnerships (kreditnye tovarishchestva) — member-owned rural lenders that largely on-lend state agricultural funding. KZT 1.1 trillion of assets.
• Pawnshops (lombardy) — KZT 0.7 trillion of assets, much of it collateral taken in pledge.

None may take deposits. Every tenge an MFO lends comes from equity, bank lines, foreign lenders or bonds — which is why they are private credit counterparties, and why the largest of them are leaving the category to become banks.

Focus of This Report

1. Kazakhstan's macro backdrop, and why aggregate credit depth hides a household leverage problem.
2. The real size and composition of the non-bank sector once reclassifications are stripped out.
3. Pricing under a 46% cap shared with banks, and what it does to MFO economics.
4. The regulatory framework, centred on the 10% overdue limit inside capital adequacy.
5. Credit quality, major players and the funding stack.
6. Risks, opportunities and entry routes for private credit investors.

Key Takeaways

• Three numbers answer "how did Kazakh microfinance grow in 2025?" The full ООМФД portfolio grew 12.3%; the MFO-only series fell 4.0% to KZT 1,501 billion; and MFOs excluding the two that became banks grew 20.3%. Only the last describes the market an investor can lend into.
• Car finance now leads. Toyota Financial Services Kazakhstan, MyCar Finance and Shinhan Finance hold 39.4% of the MFO loan book, up from 30.7% a year earlier, helped by an exemption of car-secured microloans from the debt-service cap that expires on 31 December 2026. The 18 MFOs with assets above KZT 10 billion control 83.7% of sector assets.
• The headline NPL ratio is falling while early arrears explode. In the first half of 2026 loans 1–30 days overdue rose 57% to KZT 145.7 billion, total overdue rose to 23.3% of the MFO book, and the 90+ bucket fell 6%. That divergence is what a regulatory ceiling on 90+ arrears looks like from the outside.
• Profitability halved. MFO net profit fell 50.6% in 2025 to KZT 49 billion and a further 42.7% year on year in Q1 2026. Part of that is the exit of KMF and BNK; like-for-like, 2025 profit still fell 44%, on a 10% limit on insurance commissions and heavier provisioning.
• Monetary policy is now easing. The base rate has fallen from 18% to 16.25% since June 2026, with inflation at 9.3% in September. The 46% cap does not move with it, so lender spreads widen — unless the cap is cut, which the NBK governor said in October 2025 was under consideration.
• The graduation path to a banking licence is now statutory. The new Law on Banks, in force since 19 March 2026, created a basic banking licence (KZT 10 billion of capital, assets capped at KZT 500 billion) designed for large MFOs. KMF became a bank in August 2025 and now pays about 16% on deposits; MFOs fund with tenge bonds at 20–25%.
• The collector-sale moratorium expires on 1 May 2027. If it lapses, a large backlog of non-performing consumer paper becomes saleable. If it is extended again, the closed loop stays closed.

The Big Picture: Kazakhstan's Macro and Credit Overview (2025–2026)

Growth Without Oil, Inflation Coming Down

The economy grew 6.5% in 2025, then slowed to 4.2% in January–September 2026 as oil output fell 8.3% to 69.4 million tonnes. The non-oil economy did the work: construction, manufacturing, transport and trade led growth in the first half. The National Bank of Kazakhstan (NBK) forecasts 4.5–5.5% for the full year; S&P expects 5.1%.

The more important turn is in prices. Inflation fell to 9.3% in September 2026 from 12.9% a year earlier — twelve consecutive months of disinflation — and the NBK has cut its base rate three times, from 18% to 16.25%. The next decision is due on 23 October 2026.

Table 1 — Core Macroeconomic Indicators: Kazakhstan

Sources: Bureau of National Statistics; Ministry of National Economy; NBK; World Bank Kazakhstan Economic Update (February 2026); IMF World Economic Outlook (April 2026); Moody's (29 September 2026), S&P (21 August 2026), Fitch (June 2026).

Monetary Conditions and the Cost of Money

Table 2 — Monetary and Funding Indicators

Sources: NBK; ARDFM–NBK joint resolution of 28 April 2026; KASE placement data via Halyk Finance (April 2026); Finratings.kz deposit and lending rate data (2026).

Investor implication. Kazakhstan's cap is fixed while its policy rate is falling. Every 100 basis points of easing widens the gap between an MFO's funding cost and the most it can charge. That is good for lenders only as long as the regulators leave the cap alone — and in late October 2025 NBK Governor Timur Suleimenov said a further cut to 40% was under consideration.

Credit Depth: Under-Levered Companies, Over-Levered Households

On the standard metric Kazakhstan looks under-penetrated. Private credit has hovered around a quarter of GDP — below Uzbekistan and Mongolia, as earlier Kilde analyses show. That number is mostly a statement about corporate credit. The household picture is the opposite.

Table 3 — Household Leverage

Sources: World Bank Kazakhstan Economic Update, February 2026; First Credit Bureau; ARDFM; NBK. GDP ratios derived from bank loans to the economy of KZT 40.0 trillion (January 2026), household debt of KZT 27.5 trillion and 2025 GDP of about KZT 155 trillion (Bureau of National Statistics).

The World Bank's reading is blunt: households are increasingly using credit to replace income rather than to add to it, at a time when real incomes fell in 2025. Read against that, Kazakhstan is not an underbanked consumer market waiting to be served. It is a leveraged one that the regulator is actively de-leveraging, borrower by borrower. The binding constraint on MFO growth is borrower eligibility, not demand — and the eligibility rules tighten again in 2027.

Structure and Size of Kazakhstan's Non-Bank Sector

The Definitional Trap: Three Answers to One Question

Kazakh statistics publish at least three versions of the microfinance market, and they moved in opposite directions in 2025.

Table 4 — Three Measures of 2025 Growth

Sources: ARDFM, state of microfinance organisations at 1 January 2026; NBK microfinance statistics via Halyk Finance (April 2026); Kursiv Research calculations on NBK data (March 2026).

The fall in the middle row is not a contraction. When KMF, historically the largest MFO, received its banking licence on 12 August 2025, KZT 306.2 billion left the MFO series in a single month. BNK Finance Kazakhstan, a subsidiary of South Korea's BNK Financial Group, had converted in June. ARDFM puts the two converted MFOs at 37.7% of the microfinance market in its business-lending series. Halyk Finance estimates that without the KMF effect the MFO book would have ended 2025 at about KZT 1.81 trillion.

The top row is misleading in a different way. Credit partnerships — KZT 1.1 trillion of assets — are seasonal agricultural lenders: the state's Agrarian Credit Corporation alone had KZT 342 billion lent to 207 of them at the start of 2026. Their loans are repaid after harvest, which is why the ООМФД business book dips in the first half of each year. They are not a channel a foreign private credit investor can realistically enter.

Table 5 — ООМФД Composition at 1 July 2026

Source: ARDFM, report on organisations engaged in microfinance activity at 1 July 2026.

The MFO Series, Properly Read

Table 6 — MFO Loan Portfolio (KZT billion, 1 January each year)

Source: NBK microfinance statistics via Halyk Finance (April 2026); 1 July 2026 derived from NBK overdue data published by Finratings.kz (KZT 368 billion overdue = 23.3% of portfolio). The 2026 figure excludes KMF from September 2025.

The book grew 3.5× in four years, then stalled. Underneath the flat line the composition is changing quickly:

• Fewer, larger loans. The number of individual MFO borrowers fell 17.2% in the year to May 2026, to 1.5 million, and active contracts fell 22.3% to 3.1 million, while the portfolio held roughly flat. Average balances are rising as car finance gains share and PDL volumes are squeezed.
• Car finance has replaced business microcredit. Toyota, MyCar and Shinhan hold 39.4% of the loan book against 30.7% a year earlier. Regulation has helped: car-secured microloans for a new car, or a used car with a 30–50% down payment, are exempt from the 0.5 debt-service cap until 31 December 2026, and car-secured consumer loans carry a 100% risk weight in the capital ratio against 200% for unsecured ones. ARDFM notes that "classic" MFOs and car-finance MFOs are now roughly equal by assets, and only three large MFOs still lend meaningfully to businesses: Solva (OnlineKazFinance), Arnur Credit and Asian Credit Fund.
• Concentration is high. The top ten MFOs hold more than 73% of sector assets. The 18 MFOs above KZT 10 billion of assets hold 83.7% of assets and 83.1% of loans.
• Turnover in the licence base. In 2025, 17 MFOs were licensed and 18 had licences revoked.

Where Banks Sit

Banks dominate. Their KZT 25.8 trillion book to individuals is more than eighteen times the ООМФД book to individuals, and MFO lending to individuals — about KZT 1.33 trillion — is roughly 5% of total household debt. Banks are also the price-setters: they face the same 46% unsecured cap, fund at deposit rates of about 16%, and lend to prime borrowers at around 21%.

For investors: Kazakhstan's MFOs are not competing on access, they are competing on eligibility. They lend to customers banks have declined or not reached, under borrower-level rules that apply to banks and MFOs alike. Growth now comes from share — mostly in car finance — rather than from new borrowers.

Pricing: One Cap for Everyone

Kazakhstan sets its consumer credit ceiling by joint resolution of ARDFM and the NBK, re-issued for fixed periods rather than moved mechanically. The current resolution, adopted in April 2026 to align the caps with the new Law on Banks, took effect on 16 May 2026 and was amended from 1 July 2026 to introduce LTV-based mortgage caps. Kilde examined the 2024–25 tightening and its enforcement in a separate analysis of Kazakhstan's interest rate cap; the point here is what the cap does to lender economics.

Table 7 — Maximum Annual Effective Rates (AEIR)

Sources: ARDFM–NBK joint resolutions of 16/19 August 2024 and April 2026; NBK information notices of 15 May and 1 July 2026; Zakon.kz on the pre-2024 PDL limits.

Three features matter more than the headline number:

1. It is all-in. Every payment the borrower must make to obtain the loan — interest, fees, mandatory insurance, valuation, guarantor and intermediary charges — counts toward the 46%. Since August 2025 lenders acting as insurance agents can keep no more than 10% of the premium as commission, against up to 90% before. ARDFM attributes much of the 2025 fall in MFO revenue to that single change.
2. It is shared with banks. An MFO has no regulatory pricing advantage over Kazakhstan's banks. It can only win on speed, reach or risk appetite — and risk appetite is exactly what the eligibility rules limit.
3. It is fixed while rates fall. The gap between the 46% ceiling and the base rate has widened from 28 to almost 30 percentage points since May.

The Funding Arithmetic Behind Bank Conversion

An MFO lending unsecured at the cap and funding through the local bond market at 22–23% has a gross spread of roughly 23–24 percentage points before credit losses and operating costs. The same loan funded with 16% deposits earns about 30 points. That six-to-seven-point gap is the business case for the basic banking licence — and the reason KMF, BNK Finance and Solva all chose, or are pursuing, conversion rather than staying non-bank.

Tax narrows the gap slightly. From 2026 banks pay 25% corporate income tax on retail lending, five points above the general 20% rate, so a converting MFO hands part of its funding gain back in tax.

It also marks the risk in the other direction. A cut in the cap from 46% to 40%, which the NBK governor said in October 2025 was under consideration, would remove almost exactly the funding advantage an MFO gains by becoming a bank. For a lender that cannot convert, it would remove a quarter of its gross spread.

{{cta-component}}

Regulatory Framework for Non-Bank Lenders in Kazakhstan

1. Overview

2. The 10% Ceiling Inside the Capital Ratio

Since January 2023 an MFO's compliance with its capital adequacy ratio k1 — equity of at least 10% of risk-weighted assets — has been conditional on a second test: the share of microcredits more than 90 days overdue in the total loan portfolio. The limit was set at 20%. ARDFM Board Resolution No. 65 of 16 October 2025 cut it to 10%, applicable from 1 January 2026.

The definition is the mechanism. Both the numerator and the denominator include loans the MFO holds on balance sheet and loans it has written off to off-balance-sheet accounts. A write-off does not improve the ratio. A breach is a breach of k1, and the MFO must submit a remediation plan to ARDFM within one business day.

That leaves an MFO four ways to manage the ratio: underwrite better, collect, restructure — or grow the denominator.

The risk weights reinforce the message. Unsecured consumer microloans carry a 200% weight in k1 and PDL loans 500%, so a payday book needs roughly two and a half times the capital of an ordinary consumer book of the same size.

3. Every Other Exit Is Closed

• No sale. Since 1 July 2024 banks and MFOs may not sell consumer loans to collection agencies. The moratorium, originally due to end on 1 May 2026, was extended to 1 May 2027 in the January 2026 legislative package that accompanied the new Law on Banks (Law No. 259-VIII). Collectors may still work the debt as servicers, without the creditor changing.
• No refinancing. Since 1 July 2024 no new consumer loan may go to a borrower with 90+ day arrears. Since 24 November 2025 an MFO may not lend to anyone more than 30 days late on a bank loan or more than one day late on a microloan, and the debt-service cap for borrowers with 90+ day arrears in the previous year was halved from 0.5 to 0.25.
• No serial restructuring. A borrower whose debt was fully forgiven in the past 36 months, or who had a restructuring in the past 12 months that failed — the balance did not fall by 5% within six months, there were two or more restructurings with arrears, or arrears were refinanced with a new loan — cannot be lent to either.
• No long tenors. Unsecured microloans longer than five years are prohibited; three-to-five-year loans are classed as high-risk and carry heavier provisions.
• No escalating PDL costs. PDL extensions are capped at 90 days, and penalties, fines and fees may not exceed half of the original principal.

4. Personal Bankruptcy

The Law on Restoration of Solvency and Bankruptcy of Citizens, in force since 3 March 2023, gives borrowers an extrajudicial route that covers debts to banks, MFOs and collection agencies. More than 270,000 applications have been filed. In 2025 more than 51,000 people were declared bankrupt and KZT 188 billion of debt was written off. Creditors cannot initiate the procedure. Amendments in 2026 widened access further — the Ministry of Finance estimates around 300,000 people could become eligible — and removed pawnshops from the extrajudicial route.

For an unsecured lender this is a hard floor on recoveries: a borrower who qualifies exits with the debt extinguished.

5. Behavioural Supervision — and the Restructuring Problem

ARDFM supervises each MFO's problem-loan reduction plan monthly. In 2025, MFOs and collectors gave 722,000 borrowers owing KZT 399 billion easier repayment schedules, and wrote off the debts of 34,000 borrowers (KZT 2.2 billion) entirely.

ARDFM's own assessment is unusually candid: most of these measures were short technical deferrals granted without assessing the borrower's finances, and much of the debt fell back into arrears. The regulator now classes ten MFOs, unnamed, as high credit risk — weak portfolios, poor affordability checks, thin reserve coverage and weak capital. It has announced a new supervisory model built on mandatory self-regulation, behavioural scoring of lenders and product-governance rules covering needs assessment, advertising and product changes.

6. The Banking Ladder

The new Law "On Banks and Banking Activity" (No. 258-VIII of 16 January 2026, in force 19 March 2026) introduced two licence tiers. The basic licence requires KZT 10 billion of capital, caps assets at KZT 500 billion and excludes lending to non-residents and higher-risk securities activity. ARDFM has said it is designed for converting MFOs. The universal licence requires KZT 20 billion.

7. What Changes Next

Investor implication: Kazakhstan's rules are tight, explicit and enforced. ARDFM revokes licences, suspends lenders that ignore its orders and publishes its supervisory priorities. The regulatory risk is not arbitrariness; it is direction. Every change since 2019 has moved the same way, and a credit case that depends on the 46% cap, the 2027 moratorium expiry or the 10% limit staying where they are should say so explicitly.

Credit Quality: The Pipe Is Filling at the Top

Table 8 — MFO Overdue Loans by Bucket (KZT billion)

Source: NBK microfinance statistics, published by Finratings.kz (17 August 2026). The 31–90 day row for 1 January 2026 is derived as total less the published 1–30 and 90+ buckets.

Read the two bold rows together. Early arrears rose 57% in six months while the 90+ bucket shrank. Individuals account for 88% of all overdue MFO debt, and their arrears grew 15.1% in the half-year against portfolio growth of 3.1%. Sole proprietors are worse: 29.1% of that book is overdue.

This is the signature of a ceiling on one bucket. Loans approaching 90 days are restructured, rescheduled or pushed to cure — which ARDFM itself says often does not hold — while the borrower-level gates prevent the customer from refinancing elsewhere. The result is a 90+ ratio that flatters the book and early buckets that keep refilling.

Table 9 — The 2025 Path

Source: NBK microfinance statistics via Halyk Finance (April 2026).

Two things the headline hides. First, reserve coverage of 90+ loans fell from 85.5% to 68.0% during 2025 even as the 90+ ratio improved — less protection behind a ratio that is being managed. Second, the NPL 90+ figure published in the NBK series is on-balance-sheet only. The regulatory ratio that decides k1 compliance adds back written-off loans, so it is higher. Lender-level regulatory ratios are not public.

The Dispersion Is the Data Point

Source: Kursiv Research on NBK data (26 March 2026).

The last-but-one row is the mechanism in a single lender: three-quarters of the book more than 30 days late, but less than 6% beyond 90 days.

Kilde view. In Kazakhstan, NPL 90+ is a regulated output, not a measured input. We underwrite Kazakh lenders on roll rates out of the 1–30 day bucket, on the re-default rate of restructured loans and on the regulatory overdue ratio including write-offs — and we write facility covenants on the same definition the regulator uses, so that covenant headroom and k1 headroom are measured in the same units. A lender that cannot produce those three numbers monthly is not ready for institutional capital.

Major Players and Market Map

Table 10 — Largest MFOs at 1 January 2026

Sources: NBK microfinance statistics and KASE data via Halyk Finance (10 April 2026); KASE issuer pages; Fitch rating releases for Arnur Credit (29 June 2026), TAS Finance Group (23 April 2026), Jet Finance (April 2026) and MyCar Finance (August 2026). Total MFO assets were about KZT 1.7 trillion at 1 January 2026.

What the Table Shows

• Toyota is now the largest MFO by a wide margin, at almost 19% of sector assets, with the best asset quality and an investment-grade rating driven by its parent. It added KZT 60 billion of loans in 2025 and would rank around 15th among Kazakhstan's banks by loan book.
• The largest universal MFO is under strain. Solva reported the sector's worst profit swing in 2025 (KZT −11.2 billion against 2024), attributed it to additional provisioning ahead of its planned conversion to a bank, and in March 2026 publicly denied market rumours of a default, stating that its bonds are serviced on schedule.
• Profitability and asset quality do not line up. The most profitable lenders by ROE — Jet Finance and Robocash.kz — fund in US dollars on KASE and run PDL or car-secured books. Robocash.kz has a 5.4% NPL 90+ ratio and a 47.2% NPL 30+ ratio at the same time.
• Enforcement reaches the largest earners. Robocash.kz, owned by Singapore-based Robocash Pte Ltd, had its lending suspended for two months in September 2025 for failing to comply with a regulator's written order. Its loan book shrank by about 30% in 2025, and Fitch has its B− rating on negative outlook.
• Sub-scale tail. Below the top 18, nearly 200 MFOs share about 16% of assets. Several smaller KASE issuers report NPL 90+ above 8% and negative returns.

Table 11 — Business Models and Private Credit Relevance

The Funding Stack

MFOs cannot take deposits. Borrowed funds make up 61.5% of ООМФД liabilities (KZT 1.4 trillion), down from 69.5% at the start of 2025 as equity has built up.

The currency read. The tenge strengthened by about 11% in 2026, from 505.73 to roughly 451 per dollar by mid-October, helped by oil prices and non-resident inflows into government bonds. That has flattered every MFO that borrowed in dollars to lend in tenge. It is not a base case to extend: in September 2026 professional market participants surveyed by the Association of Financiers of Kazakhstan expected the dollar above KZT 511 within a year, and the Eurasian Development Bank had projected about 544 for end-2026 earlier in the year. A dollar facility to a tenge lender needs either a hedge or enough margin to absorb a 15–20% move.

The tax read. Interest paid to a foreign lender bears 15% Kazakh withholding tax. For a Singapore-resident beneficial owner the 2006 Singapore–Kazakhstan treaty reduces it to 10%. Either way it is a cost that must be grossed up or priced.

The retail read. Kazakh MFO bonds are, increasingly, a retail product priced above 20%. A market that deep in retail money is a reliable source of funding while sentiment holds — and a fragile one when a large issuer faces a rumour, as Solva did in March 2026.

Kilde view. The basic banking licence changes what a facility to a large Kazakh MFO is. For a lender on the conversion path it is bridge financing: deposits at around 16% will replace any private credit line priced above that the moment the licence arrives. We price and structure such facilities for prepayment, not for a full term. The lasting counterparties are the lenders that will stay non-bank — mid-sized, profitable, granular — for whom a committed, covenanted facility replaces retail bond refinancing risk rather than competes with deposits.

Market Risks

1. Regulatory Risks

2. Credit and Market Risks

• Early arrears are rising fast. Loans 1–30 days overdue rose 57% in the first half of 2026; total overdue reached 23.3% of the MFO book.
• Restructuring masks stress. KZT 399 billion was rescheduled in 2025 for 722,000 borrowers, a sum equal to about 30% of today's MFO book to individuals — although part of it sat with collectors — and ARDFM reports widespread re-default.
• Households are stretched. Debt at 51% of wages, four loans per borrower and falling real incomes in 2025.
• Bankruptcy is widening. KZT 188 billion of personal debt was extinguished in 2025, and eligibility expanded again in 2026.

3. Funding Risks

• No deposits. Permanent wholesale and retail-bond dependence at 20–25% in tenge.
• Currency. Dollar funding into tenge books after an 11% tenge rally.
• Graduation prepayment. Lenders that obtain a basic licence refinance with deposits.

4. Outlook

• Short term: NPL 30+ keeps rising through 2026 as tightened eligibility rules bite; further licence revocations among the ten high-risk MFOs; spreads widen if easing continues.
• Medium term: two or three more MFO-to-bank conversions; the 1 May 2027 moratorium decision sets the NPL market's direction.
• Long term: a smaller sector led by car finance and a handful of universal lenders, regulated almost as tightly as banks.

The falsifier. The thesis of this report is that the 10% limit and the sale ban make the reported 90+ ratio a managed number, and that real stress shows first in early buckets and restructured loans. If the 1–30 day bucket stops growing and restructured loans hold without re-defaulting through H2 2026, the stress reading was too pessimistic and the sector is simply digesting tighter rules. If, instead, several mid-sized MFOs breach the 10% limit or lose licences before the moratorium decision, the closed loop is producing losses faster than lenders can absorb them.

{{cta-component}}

Opportunities for Private Credit Investors in Kazakhstan

Kazakhstan combines an investment-grade sovereign, a strong and active regulator, a functioning local bond market and lender disclosure published monthly by the central bank. It also combines a capped price, a constrained borrower base and a regulatory design that keeps bad debt on lenders' books.

The Five Routes In

Indicative Return Spectrum

Directional, not quoted. Pricing varies with security package, tenor, currency, hedging and issuer scale. Any yield quoted without a matching security and covenant description is not a usable number.

Mitigating Factors Specific to Kazakhstan

• Investment-grade sovereign with large external buffers, rated BBB/Baa1 by all three agencies.
• Granular public data: the NBK publishes MFO-level portfolio, arrears and profitability data, and ARDFM publishes sector reports quarterly.
• Rated, listed counterparties: most large MFOs file with KASE, and several carry Fitch or S&P ratings.
• A predictable regulator whose direction of travel, while restrictive, is consistently signalled in advance.
• A strong credit bureau system that underpins the borrower-level rules.
• A securitisation framework for MFOs, under which pools of loans no more than seven days overdue can be moved to a special purpose company with capital relief — a ready-made structure for an asset-backed facility.

How Kazakhstan Compares

Table 12 — Kazakhstan in Context

Sources: Kilde market analyses (2025–2026); NBK; ARDFM. Figures are the latest available and not always at the same date.

The rest of the series, including Poland, Spain, India and the Philippines, is in Kilde's market reports.

Interpretation. Uzbekistan offers growth without a cap and Mongolia scale with a high one. Estonia offers a legal wrapper. Kazakhstan offers something different: the most institutionally mature market in Central Asia, with a regulator that has deliberately traded lender returns for borrower protection and has written that trade into the capital rules. It rewards investors who can underwrite to the regulatory definitions — and punishes those who read the published NPL ratio at face value.

Conclusion

Kazakhstan's licensed non-bank sector is larger, better capitalised and more tightly regulated than its headline numbers suggest.

The numbers need reading properly. The MFO book did not shrink 4% in 2025; KMF and BNK Finance became banks, and the rest grew about 20%, led by car finance — three car lenders now hold almost 40% of MFO loans. Private credit is a quarter of GDP, but household debt is at a record 51% of wages, and the regulator is de-leveraging borrowers one rule at a time. Profitability halved under a 46% all-in cap that applies to banks and MFOs alike.

The distinctive risk is the closed loop. Since 1 January 2026 an MFO's capital ratio depends on keeping 90+ day arrears — including write-offs — below 10% of its portfolio, with debt sales banned and refinancing closed to any borrower in arrears. The 90+ ratio has fallen while loans 1–30 days overdue rose 57% in six months. The published figure is the one number in the market a lender is required to manage.

The distinctive opportunity is the same structure seen from the lender's side. Mid-sized, profitable MFOs that will remain non-bank need committed funding to replace a retail bond market priced above 20%, and a facility built on the regulator's own definitions gives an investor a better view of credit quality than any public ratio.

Kazakhstan does not reward investors looking for an underbanked consumer market. It rewards investors who can underwrite the early buckets, price the regulator's next move — and who understand that in a market where bad debt cannot leave the building, the most important number is the one nobody publishes.

About Kilde

At Kilde, we specialise in secured private credit — and that is all we do.

Kilde is a regulated investment platform that helps accredited investors earn a predictable monthly income through high-yield, senior-secured private credit. We deal exclusively with licensed non-bank financial institutions with a proven track record and strong financial performance. Each borrower is vetted using Kilde's proprietary credit scoring model, which analyses over 20 million data points on loans and repayments.

Our investments typically run 3 to 36 months with collateral coverage above 100%. Monthly cash coupons provide predictable income, and an early redemption option on select deals offers additional flexibility. Strict covenants on leverage, interest coverage and repayment rates protect investor capital throughout the tenure, with continuous monitoring of borrower performance and loan agreements under Singapore law.

Where conventional fixed income forces a choice between low-risk, low-return safety and higher-risk vehicles, Kilde's model delivers strong yields with disciplined risk management.

Sources

{{sources}}