BRAC Bank valuation report: fair value BDT 18.4–73.3 (Oct 2026)

Fair value BDT 18–73: the price of 67.5 assumes the treasury windfall lasts

Updated 31-Oct-2026: added the BRAC Bank green bond (BDT 10bn, zero-coupon, up to three years; not treated as Tier-2 capital) and Bangladesh Bank's May-2026 dividend rule. BRAC clears the BDT 20bn paid-up line, but the cash-dividend caps now limit our forecast cash dividend to BDT 1.50 a share, which moves the base to 41.5 from 42.2. Prices and filings are unchanged (price 29-Oct-2026, data to 30-Sep-2026).

We put fair value at BDT 18–73, base 42 (probability-weighted 43.7), against a price of 67.5. BRAC Bank earned BDT 15.8bn standalone in FY2025, up 30% [1 p.163], and BDT 15.8bn in the nine months to September 2026, up 50% y/y [4 p.23]. Reported net interest income fell from BDT 17.7bn in FY2023 to 11.8bn in FY2025 while investment income rose from 7.6bn to 37.9bn [1 p.162], but that compares a net figure with a gross one: the deposits that fund the BDT 326bn bond book are charged inside interest expense. Charging the bond book its share of funding cost, treasury contributed about BDT 19.8bn net, 34% of FY2025 operating income, and lending about 30.0bn (model, [8]).

Our explicit forecast lets the windfall fade slowly: investment yields fall from 12.3% in FY2026F to 10.5% by FY2028F, and net profit grows 9% a year from FY2026F to BDT 25.3bn in FY2028F (9.1% in FY2027F). For the long run we go further, as our bank method requires: we reset the bond yield to the T-bond less 1pp (8.27%), let deposit costs follow 30% of that fall, and use the five-year average credit cost. That normalised ROE is 13.8%, against 15.6% in FY2028F. At a 17.2% cost of equity it justifies about 0.67x adjusted book; in our sum of the parts the bank comes to BDT 31.3 a share, bKash 7.8 and the other subsidiaries 0.8.

The price implies a sustainable ROE of about 21.7% on our cost of equity, against 18.0% in FY2025 and a five-year median of 12.0% [1 p.164]. So the market assumes the bank keeps its treasury-era profitability for good. Asset quality is not the issue yet: gross NPL was 2.27% in September 2026 and provisions exceeded the requirement by BDT 2.3bn [4 p.16].

We rate uncertainty High rather than Medium. The range is wide, BDT 18 to 73, because two inputs that nobody can observe (how fast treasury yields fall and what return the bank earns afterwards) drive most of the value. The stock rose 76% in two years against 3% for the DSEX [5], so the market has already rewarded the earnings jump.

BRAC Bank built a BDT 1.18trn balance sheet on SME lending: loans 2.3x and deposits 2.8x in four years

Standalone total assets reached BDT 1,179.7bn at end-2025, from 449.1bn at end-2021 [1 p.162]. Loans grew from BDT 321.2bn to 730.7bn and deposits from 314.6bn to 889.5bn over the same four years. By September 2026 deposits were BDT 1,030.9bn and loans 779.5bn [4 p.7]. Deposits have outgrown loans every year since FY2023, which is why the bank now holds BDT 427.2bn of investments, 31% of assets [4 p.7].

SME lending is the core of the loan book: BDT 306.2bn, 42% of loans at end-2025, down from 47% in FY2021 as retail and corporate lending grew faster [1 p.163]. The network is 194 branches, 116 sub-branches (up from 74 a year earlier), 446 SME unit offices and 329 ATMs, serving 2.3mn customers with 10,308 staff [1 p.165]. Inward remittance handled reached BDT 373.0bn in FY2025, up 29% [1 p.163], which feeds both deposits and FX income.

The group has four subsidiaries and one associate. bKash, the mobile financial service, is the only one that matters for value: BDT 6.8bn profit in FY2025 on BDT 65.7bn revenue, with BDT 110.0bn of e-money in circulation [s3 p.4–5]. BRAC Bank owns 51% of its ordinary shares, but its economic share is 35.74% because of bKash's preference shares [1 p.508]. The two capital-market subsidiaries and the UK remittance company together reported a small loss in FY2025 [1 p.135].

The other subsidiaries are small: BRAC EPL Investments (99.95%, profit BDT 18mn), BRAC EPL Stock Brokerage (90%, 67mn), BRAC Saajan Exchange in the UK (98.08%, a loss) and the associate BRAC IT Services; together they are carried at BDT 2.3bn [s1 p.135, p.508].

In the standalone accounts BRAC carries bKash at its share of net assets, BDT 14.8bn, with the fair-value gain in equity [1 p.508]. bKash profit therefore never reaches standalone earnings. We value the bank on standalone numbers and add bKash separately, which avoids double counting.

bKash: BDT 7.1bn of FY2026E profit and 7.8 a share of value, with Bangla QR the new competitive test

bKash is Bangladesh's largest mobile financial service: 82.2mn customers (57.5% active in 90 days), about 380,000 agents and BDT 13.97trn of transactions in 2025, up 33% [s1 p.586–587, p.590]. It earned BDT 6.8bn after tax on BDT 65.7bn revenue in FY2025, up from 3.2bn, and held BDT 41.3bn of equity and BDT 110.0bn of customer e-money [s3 p.4–5]. Agent cash-in and cash-out still make up most of the volume [1 p.587].

bKash's 2026 results are not published separately, so we derive them from BRAC's consolidated accounts. Non-controlling interests, almost all of them bKash's other shareholders (64.26%), took BDT 3.58bn of 9M 2026 profit against 3.24bn a year earlier [4 p.4]. That implies bKash earned about BDT 5.6bn in 9M 2026, up about 10%; the same method gives 6.77bn for FY2025 against the reported 6.76bn. Adding Q4 at the Q3 2026 run-rate (BDT 1.48bn, not grown), we use FY2026E profit of BDT 7.1bn. The Q3 non-controlling share fell to BDT 0.95bn from 1.27bn a year earlier [4 p.4], so growth slowed in the third quarter.

Bangla QR is the main new competitive test. Bangladesh Bank's 2021 guideline set Bangla QR as the national interoperable QR standard: any bank account, card, MFS or e-wallet can pay any merchant's code, proprietary codes were to be replaced, and acquirers may not pass the merchant discount rate on to customers [s12 p.1, p.9]. In April 2026 BB made it mandatory at all merchant points, with proprietary QR codes to be replaced by 30 June 2026, under the Payment and Settlement Systems Act 2024; bKash said it would comply [s13; s14]. The Daily Star reported, citing BB data, that daily Bangla QR transactions rose from about 1.78 lakh to 3 lakh and merchants from about 16 lakh to over 30 lakh after the July rollout [15].

Interoperability weakens one part of bKash's moat: a merchant that took bKash QR because customers used bKash can now accept any wallet or bank app on one code, so banks and other wallets can compete for the same payment. The evidence on how much this costs bKash is thin: bKash does not disclose merchant-payment revenue, and we found no official merchant-fee data. We therefore make it a stated haircut to bKash profit rather than a forecast: 5% in the base case, 15% in the bear case, none in the bull. The bear haircut alone takes BDT 0.8 off the bKash part and 1.6% off base fair value. Cash-in/out through agents, which is most of the volume, is not affected by Bangla QR.

A 2021 transaction gives context, not a price for today. SoftBank Vision Fund II agreed to buy a 20% stake in bKash through new and existing shares in November 2021 [s10; s11]. The Business Standard alone reported, citing BRAC Bank's chairman, that the deal would value bKash at about US$2bn, which it put at more than Tk17,000 crore (BDT 170bn) [10]. That is well over twice our base value for 100% of bKash, about BDT 67bn. We do not use it as today's price: a strategic buyer paid for a 20% block and future growth when bKash was loss-making (a BDT 1.2bn loss in 2021 [1 p.590]); growth is now slowing (the Q3 2026 non-controlling share was 25% below Q3 2025 [4 p.4]); BRAC holds a minority economic interest in an unlisted company, hence our holding discount; and the valuation comes from a single report with no filing behind it.

The 10x multiple is a judgement. We found no citable multiples for listed mobile-money peers in our sources, so it rests only on bKash's own numbers: profit growing about 10% in 2026 after more than doubling in 2025, and a business still dominated by agent cash-in and cash-out. The sensitivity below shows what other multiples would do.

Deposits grew 29% in FY2025 and CASA held at 42%: funding is not the constraint

BRAC Bank added BDT 200.1bn of deposits in FY2025 (+29.0%) and 141.4bn in the nine months to September 2026 (+15.9%) [s1 p.162; s4 p.7]. Current and savings accounts were 41.9% of deposits in September 2026, against 41.1% in December 2025 [4 p.7]. Cost of funds rose from 2.21% in FY2021 to 6.85% in FY2025 as rates became market-based [1 p.164]; in 9M 2026 it eased to about 6.4% annualised (our calculation from [s4 p.7, p.9]).

Liquidity is far above every requirement. The statutory liquidity ratio was 39.78% in September 2026 against 13% required [4 p.14], and loans were 75.6% of deposits [4 p.7]. The liquidity coverage ratio was 291.95% at end-2025 [1 p.162]. Borrowings from the central bank and government agencies, which are refinance lines rather than emergency support, fell from BDT 52.0bn to 41.6bn in nine months [4 p.7].

BRAC is growing faster than the banking system. System deposits grew 10.7% and loans 4.3% in FY2025-26 (year to June) [s9 Table-102]; BRAC's deposits grew 15.9% and loans 6.7% in just nine months of 2026 [4 p.7]. Private banks as a group grew deposits 2.3% and loans 2.0% in the April–June quarter [s9 Tables 6, 39]. BRAC's 41.9% current-and-savings share compares with 25.4% for private banks and 28.5% for all banks [s9 Tables 2, 6]. Its June deposits of BDT 979.7bn were 4.4% of the system's BDT 22.1trn [s2 p.7; s9 Table-102].

BRAC's loan yield is below the system's weighted-average lending rate (11.4% vs 12.13%) because its book includes low-risk corporate and consumer loans; its cost of funds is near the system deposit rate, despite a higher CASA share, because it also carries BDT 145bn of borrowings and bonds [4 p.7]. The money market is tight: the overnight call rate was 9.24% on 29-Oct-2026 (7.85% on 7-Oct) and the 14-day T-bill cut 8.97%, both above the 91-day bill at 8.56% [7]. An inverted front end like this is near-term pressure on deposit costs, which is why we assume only a 30% pass-through of lower bond yields.

We model deposits growing 2% in Q4 2026 (FY2026F +18%), 14% in FY2027F and 13% in FY2028F. That is well below FY2024–25, when deposits moved from weaker banks to stronger ones. Our bear case cuts growth to 9% and 8% in FY2027–28F; the FY2027F EPS effect of 5% deposit growth from FY2027 is small (BDT 9.54 vs 10.18) because surplus deposits earn little more than they cost.

Net of its funding cost, treasury earned 34% of FY2025 operating income: falling bond yields are the main earnings risk

Investment income was BDT 37.9bn in FY2025 [1 p.162], of which 31.6bn interest on government securities and 5.3bn capital gains on them [1 p.557]. Reported net interest income (11.8bn) cannot be compared with it directly: interest expense on all deposits sits inside NII, including the deposits that fund the bond book. The fall in NII since FY2023 is mostly that allocation: loans still earn a 4.2% spread over cost of funds (11.05% vs 6.85%). Charging investments their share of funding (cost of funds x average investments, BDT 18.2bn), lending NII was about BDT 30.0bn and net treasury income 19.8bn, a funded spread of 7.4% (model, [8]).

The first nine months of 2026 continued the pattern: investment income BDT 34.7bn (+32% y/y) and net interest income 11.3bn (+28%) [4 p.9]. On our numbers the annualised yield on investments was 12.3%, above the 9.27% 10-year T-bond cut-off yield of 25-Oct-2026 [7]. As older high-coupon bonds mature, the book yield will drift toward market yields. We assume 12.4% in Q4 2026, 11.2% in FY2027F and 10.5% in FY2028F.

Fee income moved the other way in FY2025. Commission, exchange and brokerage fell 14.5% to BDT 8.0bn: gains on foreign-currency dealing dropped from BDT 3.2bn to 2.5bn and loan processing fees halved from 1.7bn to 0.8bn as lending slowed [1 p.557]. The first nine months of 2026 recovered to BDT 7.0bn, up 19.5% y/y [4 p.9]. We grow fees 10% a year in FY2027–28F; at about 13% of operating income they cannot offset a treasury shortfall.

Cash conversion is adequate. Interest received in cash was BDT 104.1bn in FY2025, 92.4% of interest plus investment income, up from 89.1% in FY2024 [1 p.488]. Investment income includes revaluation and capital gains that are not interest receipts, so the true conversion of interest income is higher than this ratio.

The return on assets tree (gross presentation, as reported) shows the dependence. In FY2025 investment income was 3.60% of average assets against 1.12% from net interest income. Cost-to-income improved from 58% in FY2022 to 45% in FY2025 [1 p.164] only because treasury income swelled the denominator; operating expenses still rose 25% in FY2025.

Sensitivity is large. If investment yields are 1pp lower than our base in FY2027–28F, FY2027F EPS falls to BDT 8.90 from 10.18; at 2pp lower it is 7.61. Cheaper deposits offset part of it: each 0.5pp lower cost of funds adds about BDT 1.72 to FY2027F EPS.

Gross NPL rose to 2.27% in Q3 with 107% coverage incl. general provision: the adjusted-book haircut is only BDT 0.9 a share

Classified loans were BDT 17.7bn, 2.27% of loans, in September 2026: up from 2.03% in June but level with end-2025 and well below 3.90% at end-2021 [s4 p.16; s2 p.16; s1 p.163]. Sub-standard loans rose from BDT 2.8bn at end-2025 to 4.0bn in June and 7.0bn in September, while bad/loss loans fell from 11.1bn to 7.3bn [s4 p.16; s2 p.16]. Sub-standard is where new problem loans first appear, so the Q3 jump is the item to watch; Q3 provisions were BDT 2.6bn, against 1.4bn in Q2 [s4 p.9; s2 p.9].

Two comparisons we still cannot make: the share of rescheduled loans and the banking-system classified-loan ratio are not in our data. Bangladesh Bank's statistics give system overdue loans of 34.3% of outstanding in June 2026 [s9 Table_101], but overdue is a wider measure than classified and is not comparable with BRAC's 2.27%. The low ratio is partly the result of write-offs. In FY2025 BDT 9.4bn of loans became non-performing and 9.2bn left the category, including 7.2bn written off against provisions [1 p.445]. That is a gross slippage of about 1.4% of average loans a year: low for Bangladesh, but not zero. Recoveries of written-off loans were BDT 1.05bn [1 p.445].

Provisions exceed the Bangladesh Bank requirement. Required provision was BDT 16.6bn in September 2026 against 19.0bn maintained, an excess of 2.3bn; at end-2025 the excess was 0.4bn [4 p.16]. The required figure includes BDT 2.0bn of additional provision that BB asked for after its 2023 inspection (BDT 0.46bn in the June note) [s4 p.16; s2 p.16]. We found no provision deferral in the notes.

Our adjusted book (ABV) starts from end-2025 equity of BDT 99.7bn [1 p.162], removes the BDT 12.6bn bKash carrying value (valued separately) and deducts half of FY2026F net NPL (BDT 2.1bn). That gives BDT 82.7bn, or 36.1 a share. In the bear case the haircut doubles to 100% and NPL rises to 4.5–5.5%. Equity rose a further BDT 13.4bn by September beyond retained profit: BDT 11.8bn of revaluation gains on government securities and 1.6bn of fair-value gains on subsidiaries, mostly bKash [4 p.7]. We exclude both from the book we value: the bond gain is the present value of above-market coupons that our forecast already earns, and bKash is valued separately. On September book so adjusted, BRAC trades at 1.58x.

CRAR of 15.8% leaves BDT 22.5bn above the 12.5% minimum: dividends are affordable

Standalone CRAR was 15.76% in September 2026 with Tier-1 (all CET1) at 12.36%, against a 12.5% total requirement including the buffer [4 p.24]. The capital surplus was BDT 22.5bn. CET1 has hovered at 12.2–12.9% for five years [1 p.163]: the bank grows risk-weighted assets as fast as it retains profit, and Tier-2 capital (subordinated bonds and general provisions, BDT 23.5bn) carries the buffer.

Risk-weighted assets were 50.8% of total assets in September 2026 [4 p.24], low for a bank because government securities carry no credit risk weight. That is the other side of the treasury model: every taka moved from bonds into loans uses capital. Tier-2 capital rose to BDT 23.5bn after the social subordinated bond grew from BDT 6.5bn to 10.0bn in 2026 [s4 p.7, p.24]. We hold RWA density at 50.8% and let Tier-2 amortise to BDT 19.5bn by FY2028F; CET1 still rises to 13.1% because retained profit outpaces loan growth.

Bangladesh Bank restricts dividends at banks with high NPLs, provision shortfalls or deferrals, or capital below requirement. The circular itself is not in our document set, so we test the conditions we can check from the filings. BRAC passes all four, and the new paid-up test below. The board paid 15% cash and 15% stock for FY2025, a total payout of 38% of profit [1 p.165].

A Bangladesh Bank circular of 23 May 2026 allows cash dividends, from declarations for FY2026, only at banks with paid-up capital of at least BDT 20bn, and caps the cash part at 50% of the total dividend declared [s17; s18]. BRAC's paid-up capital is BDT 22.9bn [4 p.7], so it is above the line; The Daily Star reported it as the only listed bank that is effectively eligible [18]. The earlier rule caps the total dividend at 30% of paid-up capital and, at a CRAR of at least 15%, 50% of profit, as reported by The Daily Star [19]; BRAC's FY2025 dividend was exactly 30% of paid-up capital [1 p.165].

Those caps bind our earlier assumption. With paid-up capital held at BDT 22.9bn, the total dividend is limited to BDT 6.9bn and the cash part to half of that, BDT 3.4bn or BDT 1.50 a share, a yield of 2.2% at 67.5; our 20–25% cash-payout assumption would have paid more. We therefore cap the model's cash dividend at BDT 1.50 a share for FY2026–28F; the rest of the 30% would come as stock, which adds shares without changing value. Capital does not bind: CRAR ends FY2028F at 15.3% with BDT 25.0bn of surplus capital. Retaining more profit at a sustainable ROE below the cost of equity lowers our value slightly.

BRAC is also raising non-capital funding for green lending. The board approved a green bond of up to BDT 10bn on 21 May 2026 and BSEC gave consent on 1 October 2026: unsecured, non-convertible, fully redeemable, zero-coupon, up to three years, by private placement, with proceeds for green financing [16]. Unlike the 2025 social subordinated bond, which was raised for Tier-2 capital [16], the disclosures do not describe the green bond as Tier-2 capital, and a three-year bond is shorter than subordinated capital issues, so we leave CRAR unchanged and treat it as borrowing within our funding assumptions.

We forecast EPS of BDT 9.34 in FY2026F, rising 9% a year to 11.05 in FY2028F

Our FY2026F standalone profit is BDT 21.4bn, EPS 9.34, up 35%. Nine months are reported: BDT 15.8bn, with Q3 at 5.5bn (EPS 2.40) [s4 p.9, p.23]. We add BDT 5.5bn for Q4, about level with Q3, because Q4 carries the year-end provision true-up (we use a 1.5% annualised credit cost; Q4 2025 provisions were BDT 1.9bn [1 p.137]).

The balance sheet drives the forecast. Loans grow 9%, 12% and 13% and deposits 18%, 14% and 13% in FY2026–28F; surplus funding goes into government securities, so investments reach BDT 579bn by FY2028F. Profit grows 9.1% in FY2027F even though the investment yield drops 1.2pp, because the bond book keeps growing; it then picks up to 8.5% in FY2028F as net interest income recovers with cheaper deposits.

Costs grow faster than lending. Operating expenses rose 25% in FY2025 and 14.2% in 9M 2026 [s1 p.163; s4 p.9] as the bank added 42 sub-branches and 1,241 staff in FY2025 [1 p.165]. We grow them 13%, 12% and 11% in FY2026–28F, so cost-to-income rises from 45.0% in FY2025 to 42.4% in FY2028F. Credit cost runs at 1.07% for FY2026F (9M: 0.92% annualised) and 0.90% after, slightly above the five-year average of 0.80%. The tax rate stays near 39%, the 9M 2026 level [4 p.9].

There is no consensus forecast for Bangladeshi banks, so we compare with the bank's own run-rate and history instead. Our FY2026F EPS is close to the 9M 2026 run-rate and 35% above FY2025. Management gives no numeric profit guidance in the annual report that we could test.

The full statements, the quarterly split and every driver are in the model tabs below the article; the scenario selector on the Assumptions sheet switches all three cases.

Fair value BDT 18–73, base 42: a 13.8% normalised ROE justifies 0.67x adjusted book, plus bKash

We value BRAC Bank as a sum of the parts (75%): the standalone bank by residual income, bKash on its own earnings, and the small subsidiaries at book; a peer P/B-ROE regression (15%) and a P/E (10%) are cross-checks. Each cross-check is a whole-company value (consolidated book, or a market multiple that already prices bKash), so no part is counted twice. Residual income suits the bank: value is book plus the present value of returns above the cost of equity. Our cost of equity is 17.2%: Bangladesh Bank's 10-year bond cut-off of 9.2680% on 25-Oct-2026 [7], plus a beta of 1.06 times a 7.5% equity risk premium. We measured the beta ourselves: 1.08 on two years of weekly returns against the DSEX (our calculation from [5]), 1.06 after Blume adjustment. Our bank format's floor beta of 0.8 would give 15.3%; its typical bank range is 17–20%, and ours sits at the low end of it.

One treatment needs stating. The BDT 11.8bn rise in the bond revaluation reserve in 2026 is the market value of coupons above today's yields; our forecast earns those coupons (book yield 12.4% falling to 10.5%, against a 9.27% T-bond), so we leave the gain out of the book we charge and value, rather than count it twice. The BDT 1.6bn subsidiary fair-value gain is left out too because bKash is valued on its own. Both exclusions apply to opening book, the terminal book and the normalised ROE denominator.

Sustainable ROE is derived, not assumed. We take FY2028F, move the investment yield from 10.5% to the T-bond less 1pp (8.27%), let deposit costs follow 30% of that fall, and reset credit cost to the five-year average. Normalised profit is BDT 20.6bn and ROE 13.8%, against 15.6% in the FY2028F forecast and a five-year median of 12.0%. Because the treasury yield feeds this figure, a lower long-run bond yield lowers both EPS and fair value. The 30% deposit pass-through is a judgement within a range of roughly 20–50%. It rests mainly on BRAC's own record in a falling-yield period: in 9M 2026, against FY2025, its cost of funds fell about 20% as much as its investment yield (0.4pp vs 2.0pp, our calculation from [s4 p.7, p.9]). The upper end is the rising-rate years FY2023–25, when cost of funds followed about half of the rise in investment yield (model ratios on [1 p.162]). As context only: the system deposit rate fell 0.10pp in FY2025-26 against 0.20pp for the lending rate [s9 Table-102], and a tight call market (9.24% overnight on 29-Oct [7]) is near-term liquidity pressure rather than evidence on long-run pass-through. The choice matters: with no pass-through the base would be BDT 35.9, with our 30% it is 41.5, with 50% 45.2 and with our bank format's default 70% 48.9.

Explicit residual income over FY2026–28F has a present value of BDT 12.4bn. After FY2028F the bank earns 13.8% against a 17.2% cost of equity, so it justifies 0.67x book and the terminal term subtracts BDT 29.7bn. We remove every subsidiary from the bank's book first (bKash BDT 12.6bn net of tax, the others 2.3bn [1 p.508]), so the bank part is the bank alone: BDT 31.3 a share.

The regression leg compares like with like: screener P/B and ROE for eight profitable DSE banks are both on consolidated figures [5]. The fitted line (R² 0.21) gives 0.74x at BRAC's consolidated ROE; applied to September-2026 consolidated NAV of BDT 56.40 [4 p.23], which already includes bKash, it gives 41.9. Using the June NAV of 49.38 [2 p.3] instead would lower this leg to 36.7 and the base by about 1.3; we use September because it is now filed. Note that the September NAV includes the 2026 bond and subsidiary revaluation gains (BDT 5.85 a share) that we exclude from the residual-income leg; they add about 1.1 to the base through this leg. The P/E leg applies 8.0x, BRAC's five-year median year-end P/E on standalone EPS (13.9x, 10.8x, 7.9x, 8.0x, 8.0x for FY2021–25 [1 p.165]), to normalised EPS of BDT 6.58, not to the FY2026F peak of 9.34: those multiples were set on less treasury-dependent earnings.

At base value the stock would trade on 0.75x September-2026 standalone book and 4.4x FY2026F EPS. Solving the other way, the price of 67.5 implies a sustainable ROE of about 21.7% at our cost of equity, above even the FY2025 level of 18.0%.

Near-term asset quality moves value little. Grid B changes the NPL ratio (which feeds the book haircut) and credit cost in FY2027–28F with the long-run credit cost held at the five-year average: an 8% NPL ratio with 1.9% credit cost gives BDT 39.9. A lasting deterioration would lower sustainable ROE as well, which the bear case does.

The bear case keeps a clean book but loses the treasury windfall faster (investment yield 9.0% by FY2028F), grows deposits 8–9% and takes credit cost to 1.5%: FY2027F EPS falls to BDT 5.97 and sustainable ROE to 9.3%. It values bKash at 6x with a 15% Bangla QR haircut and a 40% discount, and the stock at today's 6.5x P/E. The bull case keeps investment yields above 11% and adds faster loan growth. All three legs are shown for each case. The bear is a severe-stress edge rather than a single likely scenario: it stacks a sustainable ROE of 9.3%, below every year from FY2021 to FY2025, a cost of equity 1pp higher, a 100% NPL haircut and bKash at 6x.

A faster fall in treasury yields is the largest single risk: 24.2% off base fair value, 25% off FY2027F EPS

Each risk below is a workbook recalculation with one input changed from the base case. Shocks that persist (bond yields, deposit costs) move fair value through normalised ROE as well as EPS; temporary ones (a two-year credit-cost spike) move EPS but barely move value, because long-run credit cost reverts to the five-year average. The bear case combines several of them with a higher cost of equity.

Two risks we cannot size well. First, regulation: Bangladesh Bank sets classification, provisioning and dividend rules, and any tightening of rules for government-securities income or dividend payouts would hit BRAC directly. Second, bond prices: BDT 11.8bn of the rise in book since December is unrealised revaluation gains on government securities [4 p.7]. We already leave it out of our valuation, but it counts in reported book and P/B and would reverse if yields rose.

Recent developments from filings and news: the auditor's FY2025 report carries an emphasis of matter on note 2, the basis of preparation under Bangladesh Bank rules on loan classification and provisioning, with an unmodified opinion [1 p.468]; the social subordinated bond rose from BDT 6.5bn to 10.0bn in 2026 [4 p.7]; and Bangla QR became mandatory at merchant points from 30 June 2026 [s13; s14]. We found no regulatory action against BRAC Bank in our sources; a bank resolution or merger programme is a sector matter we did not research for this report.

• Q4 2026: standalone profit near BDT 5.5bn keeps FY2026F on track; quarterly investment income below BDT 11bn would point to the bear case.
• FY2026 annual results (expected by April 2027): standalone EPS near BDT 9.3 and credit cost under 1%.
• Sub-standard loans: above BDT 8bn (Sep-26: 7.0bn, up from 4.0bn in June) would signal that slippage is accelerating.
• CASA share: below 40% (Sep-26: 41.9%) would mean deposit costs fall more slowly than our base case.
• FY2026 dividend (declared early 2027): our base case is 15% cash and 15% stock of face value, cash BDT 1.50 a share, the most the BB caps allow.
• Any bKash capital raise or listing: a price above 14x profit supports the bull case.

FAQ

What is BRAC Bank's fair value?

We estimate BDT 18–73 a share, with a base case of 42 and a probability-weighted value of 44, against a price of 67.5 on 29-Oct-2026. The range blends residual income on adjusted book plus the bKash stake (60%), a peer P/B-ROE regression (25%) and a P/E on normalised earnings (15%).

Why has BRAC Bank's profit grown so fast?

Mostly treasury income. Investment income rose from BDT 7.6bn in FY2023 to 37.9bn in FY2025 as deposits flowed in and were invested in high-yield government securities, while net interest income fell to 11.8bn. Nine-month 2026 standalone profit was BDT 15.8bn, up 50%.

How good is BRAC Bank's asset quality?

Gross NPL was 2.27% of loans in September 2026, up from 2.03% in June, and provisions of BDT 19.0bn exceeded the Bangladesh Bank requirement by 2.3bn. The bank wrote off BDT 7.2bn of bad loans in FY2025, which also keeps the ratio low.

Can BRAC Bank keep paying dividends?

On the tests we can check, yes: NPL well below 10%, no provision shortfall or deferral, and CRAR of 15.8% against a 12.5% minimum. We model cash DPS of BDT 1.50 for FY2026F and 1.50 for FY2027F.

How much is bKash worth to BRAC Bank shareholders?

In our base case BDT 7.8 a share in the sum of the parts (5.9 of the blended base): BRAC's 35.74% economic share of FY2026E profit of BDT 7.1bn, less a 5% allowance for Bangla QR fee pressure, at 10x and less a 25% holding discount. The bear and bull cases give 3.4 and 12.3

Sources

1. BRAC Bank PLC Annual Report 2025 (year to 31-Dec-2025), standalone and consolidated · link · accessed 2026-10-30
2. BRAC Bank PLC condensed interim financial statements, half year to 30-Jun-2026 (unaudited) · link · accessed 2026-10-30
3. bKash Limited audited financial statements 2025 (filed with BRAC Bank) · link · accessed 2026-10-30
4. BRAC Bank PLC condensed interim financial statements, nine months to 30-Sep-2026 (unaudited) · link · accessed 2026-10-30
5. AnalystBro market data: DSE prices (BRACBANK, DSEX) and bank screener, 29-Oct-2026 · link · accessed 2026-10-30
6. Bangladesh Bank call money and treasury auction pages, 29 Oct 2026 (10-year bond cut-off 9.2680% on 25-Oct-2026) · link · accessed 2026-10-30
7. The Business Standard, 'SoftBank investment to raise bKash's valuation to $2b', 14 Nov 2021 · link · accessed 2026-10-30
8. The Daily Star, 'Japan's SoftBank to buy 20pc stake in bKash', 11 Nov 2021 · link · accessed 2026-10-30
9. Bangladesh Bank PSD circular 01/2021, Guidelines for 'Bangla QR' code based payments, 6 Jan 2021 · accessed 2026-10-30
10. The Business Standard, 'Bangla QR mandatory for digital payments at all merchant points: BB', 1 Apr 2026 · link · accessed 2026-10-30
11. The Financial Express, 'BB mandates Bangla QR for all digital payments', 3 Apr 2026 · link · accessed 2026-10-30
12. DSE news, BRACBANK: Board approval to issue BRAC Bank Green Bond (23-May-2026) and BSEC's consent (1-Oct-2026) · link · accessed 2026-10-31
13. The Business Standard, 'BB bars banks with under Tk2000cr paid-up capital from declaring cash dividends', 23 May 2026 · link · accessed 2026-10-31
14. The Daily Star, 'Banks below Tk 2,000cr capital barred from cash dividends', 24 May 2026 · link · accessed 2026-10-31
15. The Daily Star, 'Many banks unlikely to pay dividends as BB tightens rules', 13 Mar 2025 · link · accessed 2026-10-31
16. The Daily Star, 'Bangla QR transactions double after mandatory rollout', 24 Aug 2026 · link · accessed 2026-10-30
17. Bangladesh Bank, Scheduled Banks Statistics Apr–Jun 2026 (Tables 2, 6, 39, 101, 102) · link · accessed 2026-10-30
18. AnalystBro model (workbook on this page) · accessed 2026-10-30