The Factory That Must Finance Itself

Executive Summary. Bangladesh has factories, entrepreneurs, banks, and a large market. What it lacks is a reliable bridge between a productive firm's present cash flow and its next machine. The World Bank Enterprise Survey found that firms financed 77.8% of fixed investment internally in 2022. Banks supplied 14.2%, supplier credit 1.1%, and equity or stock sales only 0.5%. The Bangladesh Bureau of Statistics found the same structure from the other end of the market: among 11,702,792 economic units in 2024, 9,538,495 named own funds as their main source of capital, while 321,601 named banks. These surveys cover different populations and cannot be merged, but they point in the same direction. Formal firms finance assets mainly from retained earnings, while the wider business population relies even more heavily on its own cash. Domestic credit to the private sector fell from 44.4% of GDP in 2015 to 34.5% in 2025. This is not another essay about non-performing loans or foreign investment. It is about a narrower production mechanism: when a factory must save before it can invest, viable expansion is delayed, technology adoption slows, and firm size becomes a record of past liquidity rather than future productivity.

The machine waits for yesterday's profit

A factory does not buy capacity in the same rhythm that it earns cash. A machine, testing line, boiler, warehouse, or effluent plant requires a large payment now. The revenue that justifies it arrives later, in small pieces, with risk attached. Finance exists to bridge that timing gap. When the bridge works, a firm with a credible order book can install capacity before it has accumulated the full purchase price. When the bridge fails, investment is rationed by retained earnings.

The Enterprise Survey makes that rationing visible. Internal funds covered 77.8% of fixed investment by surveyed firms. Bank finance covered 14.2%. Supplier credit covered 1.1%, equity or stock sales 0.5%, and other sources the remaining 6.4%. A reader could treat internal financing as prudence. Sometimes it is. Owners who retain profits avoid interest, dilution, and lender interference. But a system in which internal cash carries nearly the whole asset purchase is not simply conservative. It makes the timing of investment depend on the timing of profit.

Source: World Bank Enterprise Surveys, Bangladesh Country Profile, 2022.

That distinction matters because profitable firms can still be liquidity constrained. A processor may have customers, margins, and a credible reason to automate, yet lack enough accumulated cash to purchase the line. Waiting protects the balance sheet, but the delay can destroy the opportunity. A buyer changes supplier. A quality standard rises. A competitor reaches scale first. The lost output never appears as a rejected loan because the application may never be made.

This mechanism is narrower than the familiar claim that Bangladesh needs more investment. Aggregate investment can rise while the allocation channel remains weak. A large incumbent can self-finance. A protected firm can borrow against property. A new producer with a better process but fewer assets cannot easily convert expected productivity into acceptable collateral. The result is not zero factories. It is a factory population whose expansion is selected by cash history.

Credit arrives after the firm becomes large

The same survey shows a size gradient in bank access. 37.1% of small firms had a bank loan or line of credit. The share was 44.4% for medium firms and 54.6% for large firms. Across all surveyed firms it was 42.5%. Access improves as firms become larger, more documented, and more able to pledge assets.

Source: World Bank Enterprise Surveys, Bangladesh Country Profile, 2022.

There is nothing surprising about lenders preferring borrowers with records and collateral. The problem is circularity. A small firm needs finance to buy the equipment that would make it larger and easier to finance. The bank sees the current balance sheet, not the future cost curve. The firm therefore needs to cross the riskiest growth interval with the least external support.

The survey also reports that 17.2% of firms named access to finance as their biggest obstacle. That figure should not be read as the share affected by finance. It records the obstacle selected above all others. A firm that never planned an expansion because financing looked impossible may name electricity, demand, or regulation instead. A firm using internal funds may appear financially unconstrained even if that choice forced it to buy a smaller machine. Survey answers measure expressed constraint, while the financing mix reveals actual behavior.

The size gradient also changes competition. Firms with established banking relationships can respond to a large order, survive a delayed payment, or hold inventory through a disruption. Smaller rivals must turn down the order or demand cash. The advantage compounds. Scale produces bankability, bankability finances scale, and a productive newcomer can remain small without ever failing in the conventional sense.

The wider economy is even more cash-bound

The Economic Census examines a broader universe than the Enterprise Survey, including permanent and temporary establishments and household economic units. It counted 11,702,792 economic units in 2024. Among them, 9,538,495 reported own funds as the main source of capital. That is 81.5% of the total. Only 321,601, or 2.75%, named banks. 834,854 named an NGO.

These figures are not a second estimate of the Enterprise Survey indicators. The census includes tiny and household operations that the formal-firm survey excludes, and "main source of capital" is not the same measure as the percentage composition of fixed investment. Combining the denominators would be wrong. Reading the pattern across them is legitimate: dependence on own funds is visible both among surveyed formal firms and across the full business landscape.

The NGO count is also revealing. It shows that the gap is not merely between a bank loan and no finance. Bangladesh has built channels that can lend small amounts against group discipline, cash flow, or local information. Those channels support working activity, but they are poorly matched to the long life and lumpy cost of industrial equipment. A short, frequently repaid facility can keep inventory moving. It cannot safely finance a machine whose payoff arrives over years.

That mismatch helps explain why "more SME credit" is too vague a policy. Credit defined by borrower category says nothing about tenor, grace period, currency risk, collateral, or asset life. A loan can be labeled productive while its repayment schedule forces the firm to strip working capital before the machine reaches utilization. The relevant product is not any loan to a small business. It is finance shaped around the cash flow of a verified investment.

The financial system is retreating from the productive bridge

Domestic credit to the private sector was 44.4% of GDP in 2015. By 2025 it was 34.5%, a decline of 9.9 percentage points. This indicator covers more than factory credit and does not reveal sector allocation, loan quality, or maturity. It nevertheless sets the envelope inside which firms compete for finance.

Source: World Bank World Development Indicators, domestic credit to private sector, 2010-2025.

The decline matters because factory finance is not being constrained inside an expanding private-credit pool. The pool itself has shrunk relative to the economy. That does not prove that every good factory project was denied. It does mean that a narrative of abundant bank liquidity reaching productive enterprise is hard to reconcile with both the macro trend and the firm financing mix.

The policy error would be to convert this finding into a target for gross lending. Forced volume can worsen selection if banks respond by refinancing familiar clients, accepting weak collateral values, or relabeling short-term facilities. The objective is not credit for its own sake. It is a credible underwriting channel for equipment that raises measurable output, quality, energy efficiency, or export compliance.

That channel requires information the current system often does not standardize: verified invoices, machine specifications, buyer contracts, tax records, utility payments, and operating cash flow. It also requires a way to absorb first loss without hiding bad loans. A guarantee that pays whenever a politically eligible borrower defaults creates volume without discipline. A guarantee tied to independently verified equipment, lender retention of risk, and published performance can change underwriting without abolishing it.

What would change this conclusion

The conclusion would change if a new firm-level survey showed that internal finance had ceased to dominate fixed investment, bank finance had become a materially larger component, and the size gradient in loan access had narrowed without a deterioration in repayment. Until then, announced credit windows are inputs, not evidence that the productive bridge works.

Three moves would test a better model.

1. Create an equipment-finance window that pays the verified supplier, not the borrower. The facility should match repayment to the asset's operating cash flow and require the lender to retain meaningful risk. Owner: Bangladesh Bank with participating commercial banks. Success signal: the next Enterprise Survey shows bank finance above the current 14.2% share of fixed investment, with the published portfolio also reporting arrears and recoveries.
2. Turn transaction records into lendable evidence. Standardize consent-based use of tax filings, utility payments, export orders, and digital invoices for cash-flow underwriting. Owner: Bangladesh Bank, National Board of Revenue, and financial institutions. Success signal: the gap between the current 37.1% small-firm loan-access rate and the 54.6% large-firm rate narrows in a comparable survey.
3. Publish the performance of every subsidized or guaranteed SME facility. Disclose disbursement, tenor, sector, firm size, arrears, recoveries, and verified asset purchase. Owner: Bangladesh Bank's SME and Special Programmes Department. Success signal: the archive can show whether supported firms bought productive assets and repaid, rather than only how much money was announced.

The counterargument

The strongest objection is that internal finance can be a sign of strength. Bangladesh's entrepreneurs may prefer to avoid costly debt, and banks may be right to ration credit in a system where documentation is weak. Pushing loans toward small manufacturers could replace patient owner capital with fragile leverage.

That objection is valid against a volume target. It is not an answer to the timing problem. The evidence does not say every firm should borrow or that internal funds are inferior. It says external channels play a small role in fixed investment, bank access rises with firm size, and private credit has fallen relative to GDP. A good reform would preserve owner discipline while allowing a verified productive asset to be financed before the owner has saved its entire price.

The distinction is between subsidizing a borrower and underwriting an asset. The first can reward connections. The second can be checked against an invoice, installation, production record, and repayment stream. Bangladesh does not need to make factories dependent on banks. It needs to stop making a sound factory wait for yesterday's profit before buying tomorrow's machine.

---

Data sources: World Bank Enterprise Surveys Bangladesh Country Profile, BBS Economic Census Chapter 5 Excel tables, World Bank World Development Indicators, and Bangladesh Bank SME archive, all retrieved on the publication date. The Enterprise Survey covers formal private firms with five or more workers. The Economic Census covers a much wider universe. Their values are presented separately and are not pooled.

Sources

1. World Bank Enterprise Surveys, Bangladesh Country Profile 2022: https://www.enterprisesurveys.org/content/dam/enterprisesurveys/documents/country/Bangladesh-2022.pdf
2. Bangladesh Bureau of Statistics, Economic Census 2024, Chapter 5 Excel tables: https://objectstorage.ap-dcc-gazipur-1.oraclecloud15.com/n/axvjbnqprylg/b/V2Ministry/o/office-bbs/2026/4/2d39dc43-a746-4c71-b2c9-473fc81aed35.xlsx
3. World Bank WDI, Domestic credit to private sector, FS.AST.PRVT.GD.ZS: https://api.worldbank.org/v2/country/BGD/indicator/FS.AST.PRVT.GD.ZS?format=json&per_page=100
4. Bangladesh Bank, SME and Special Programmes Department archive: https://www.bb.org.bd/smespd_portal/archive_data.php