Pakistan's Local Currency Bond Market: A Narrow Investor Base, 62% of Bank Assets, and the 2028 Settlement Overhaul

On Tuesday, Pakistan's Ministry of Finance published a document called the Strategic Action Plan for the Local Currency Bond Market. Behind it sits a promise under an IMF-supported programme: identify the bottlenecks holding back rupee-denominated securities and publish a strategic plan by the end of September 2026. The promise has been kept. But three numbers in the document stopped me, and none of them is about the size of the debt. In fiscal year 2026, 91.6 percent of the government's gross borrowing of Rs34.2 trillion was raised domestically. Banks held around 78 percent of government securities. Sovereign paper accounted for roughly 62 percent of banking-system assets.

Read together, those figures describe something uncomfortable: a market with a buyer but without a market. Banks buy paper, hold it, and collect cash at maturity. The price-discovery process that gives a market its life is largely dormant. The ministry itself concedes that this concentration has supported auctions while encouraging banks to hold rather than trade, and constraining both their capacity and their incentive to lend to the private sector.

The document names the narrow investor base as the single largest gap. Pension coverage is low, insurance penetration is weak, so demand for longer-duration fixed-rate securities is thin. Foreign and retail participation remain modest. The question is not simple. When there is a captive buyer, why would you want more buyers? Because a captive buyer does not set a price; it accepts one. And in a market where prices are not discovered, the government borrows on the banks' terms, not its own.

The plan is a two-year attempt to change that. It was prepared by the Debt Management Office in the Finance Division with the State Bank of Pakistan, the Securities and Exchange Commission of Pakistan, the Pakistan Stock Exchange, the Central Depository Company and the National Clearing Company of Pakistan Limited. Its foundation is a joint IMF-World Bank diagnostic covering the money market, the primary and secondary government securities markets, the investor base, financial-market infrastructure, and the legal and regulatory framework.

Five objectives structure it: strengthening institutional capacity and coordination; making primary issuance more predictable and market-based; developing executable secondary-market liquidity and a functioning private repo market; diversifying the investor base; and modernising market infrastructure while removing legal and tax impediments. Implementation is divided into three phases: Foundations over the first twelve months; Principal market reforms between twelve and twenty-four months; and Deepening participation beyond twenty-four months.

On the primary market, the centrepiece is predictability. The government will publish target volume ranges with predefined allocation bands. Bids within announced ranges will be accepted at the market-clearing price, with deviations confined to published bands. Instrument-specific targets start with shorter maturities and expand as depth improves. Auction results will be released at a fixed time by December 2026, and a benchmark policy is targeted for June 2027.

Here lies the first important observation. A government trying to reduce uncertainty over its own borrowing cost is not the same as a government trying to reduce that cost. When auction outcomes are bounded within pre-announced ranges, investors lose some freedom to discover price. Who sets the range—the government or the market? The document does not answer that clearly, and that is the most sensitive part of the reform.

The secondary market is more transparently troubled. Trading is reasonably active up to five years of maturity, then liquidity thins. The diagnostic found the primary-dealer framework rewards turnover more than executable quotations, so a dealer can improve its score by churning paper without taking the risk of quoting prices. The framework is to be revised for FY2027/28 to give greater weight to secondary-market performance, including quote performance derived from E-Bond. A securities-lending facility for primary dealers will be assessed by September 2027, with a design and launch decision by September 2028.

For transparency, SBP and PSX will publish a daily, security-level post-trade report covering conventional securities and Sukuk. The methodology behind the Pakistan Revaluation Rates will be published, followed by a review of the yield-curve framework. Eligible bank customers will be allowed to trade exchange-listed government securities through their banks by December 2027.

Read together, these measures are essentially a plan to produce information and to make price discovery transparent—necessary for a narrow market, but not sufficient to create demand. A daily post-trade report can be published in two months; a working benchmark takes years. That gap is the real test.

On the investor base, the plan leans hardest on pension and insurance reform. The logic is simple: nobody buys long-duration paper unless they carry long-duration liabilities. Low pension coverage means few long liabilities, hence little natural demand for long fixed-rate securities. SECP's insurance and pension agendas are to be expedited, with milestones in the roadmap. But note where this sits: Phase III, beyond twenty-four months. Without it, every other reform is partial.

Retail participation is to expand through InvestPak, digital access via brokers and mutual funds, and government bond ETFs. The National Savings framework will be reviewed, covering operating costs, investment ceilings across product windows, and its interaction with the government securities market. On foreign participation, the plan explicitly references Pakistan's inclusion in the J.P. Morgan GBI-EM Edge Index, with a longer-term goal of meeting eligibility for major global local-currency indices.

Then comes the most complex engineering problem. Conventional securities settle through PRISM+, while Sukuk use infrastructure involving PSX, CDC and NCCPL. The government says this separation is not standard international practice and fragments collateral pools, limiting collateral mobility, securities lending, repo and market-making. One option is a single register for all marketable government securities operated through SBP, while preserving broker and exchange access. A decision is targeted for September 2028.

That single-register idea names a familiar modern problem: delivering ownership records, settlement finality and collateral mobility at once. It is precisely the problem distributed-ledger and blockchain-based settlement experiments elsewhere have tried to solve, with ownership and transfer in one digital record. Pakistan's plan is not blockchain-based; it is a centralised register held by SBP. That carries a real trade-off the document does not state directly. Distributed structures reduce single points of failure but raise coordination costs. Centralised structures are faster, cheaper and easier to supervise, but concentrate dependence. The September 2028 decision is not only technical; it is a decision about where power sits.

The legal and tax agenda is long and concrete: adoption of the 2026 Global Master Repurchase Agreement with Pakistan-specific provisions, or revision of the domestic master repo and netting agreements; a robust legal opinion on enforceability; SECP action on obstacles blocking non-bank repo participation, starting with money-market mutual funds; Treasury Single Account reform and stronger cash-flow forecasting; apportionment of coupon and discount income at redemption so withholding tax applies only to the final holder's accrual; aligning the tax treatment of government securities held through collective investment schemes with direct investment; completing netting legislation; and strengthening the legal basis for dematerialised holdings and settlement finality. The tax measures are targeted for the 2028-29 budget.

Governance is explicit. The steering committee is to be formed by November 2026. The detailed roadmap is due by December 2026 and will be published. The DMO will report publicly every six months through its debt bulletins. A fixed release time for auction results is due by December 2026, PKRV methodology by March 2027, and an updated DMO staffing and career framework by February 2027.

My suspicion does not come from a shortage of planning but from an abundance of it. Every reform has a date, an owner and a reporting line. Yet the market's central gap does not close by itself.

The core observation is this: Pakistan's bond market is constrained less by plumbing than by incentives. The government wants cheap, predictable funding. Banks want risk-free carry—sovereign paper, zero risk weight, a known return, cash at maturity. Both interests sustain the current concentration. If reform genuinely increases price discovery, the government must accept higher and less certain borrowing costs, and banks must receive capital incentives to trade rather than hold. Neither happens automatically.

The IMF-World Bank diagnostic frames this as a technical gap. It is also a political-economic equilibrium that is comfortable for both main participants. Breaking a comfortable equilibrium requires a crisis or an external pressure that makes reform cheaper than inertia. The 2026-27 budget and inflation management will bear directly on that.

The document's own risk list is relevant: renewed inflation, fiscal pressures, institutional capacity constraints, coordination challenges, and disruptions from liquidity, settlement and tax reforms. The ministry says phased implementation, stronger DMO capacity, cross-institutional oversight, prior consultation and regular public reporting will mitigate them. But the first two risks—inflation and fiscal pressure—are only partly within the government's control.

The retail push deserves a closer look. Bringing households into government bonds through InvestPak, digital brokers and ETFs means transferring some duration risk onto household balance sheets. Without adequate disclosure, that is not market deepening but risk redistribution. In a narrow market, more participation is not automatically better; the type of participant and their access to information matter.

Another under-discussed point: a single register under SBP increases the central bank's role in settlement infrastructure. That improves efficiency but creates a new concentration. The September 2028 decision must answer not only 'which platform' but 'which institution'.

I went back to the documents, because documents do not lie—interpretation does. The most honest sentence in this plan is probably the one conceding that concentration has supported auctions while discouraging active trading. That is the centre. Everything else—post-trade reporting, PKRV methodology, primary-dealer weighting, securities lending, the single register—is scaffolding around it.

My job is not to decide; it is to show where the decision came from. These decisions come from a narrow market that meets the government's borrowing needs on one side and shelters banks' idle balance sheets on the other. Whether reform succeeds depends on whether either side is willing to give up its comfort.

What is absent is also telling. Pension and insurance reform, the hardest lever for long-duration demand, has no specific milestone or date. The tasks that can be done quickly—publishing the roadmap, fixing the auction release time, publishing PKRV methodology, updating DMO staffing—all have firm dates. This is a familiar administrative instinct: measure what is easy first, defer what is hard.

The real metric should not be how many securities trade, but whether the spread on paper beyond five years narrows. Raising turnover is easy—securities lending or repo can temporarily lift activity. Narrowing spreads requires genuine risk-bearing capital, which comes mainly from long-duration liabilities such as pensions and insurance. The document never draws that distinction directly.

A second observation: in the three-phase structure, the first twelve months are almost entirely procedural—forming committees, adopting roadmaps, building capacity, improving communication. Necessary, but their success cannot measure market depth. The real test begins in Phases II and III, with repo documentation, securities-lending design, infrastructure architecture and tax reform. Where the obstacles are largest, the timeline is longest; where they are smallest, the promises are loudest.

A third: global index eligibility is a signal, not a cause. Inclusion in the J.P. Morgan GBI-EM Edge Index means foreign capital can buy Pakistani local-currency bonds as part of index portfolios. But inclusion does not create depth; it arrives in a market that already has liquidity, settlement finality and tax transparency. Index eligibility is an outcome, not an input.

A fourth: the split between PRISM+ for conventional securities and a separate Sukuk rail fragments collateral pools. A single register looks like a simple fix but requires coordinated resolution of broker and exchange access, settlement finality and regulatory responsibility. Otherwise a centralised register merely re-wraps an old division.

A fifth: listing 'disruptions from settlement and liquidity reform' as a risk is an odd admission. If reform itself is a risk source, its pace must be governed. Yet that governance sits with the beneficiaries and implementers themselves—DMO, SBP, SECP—without independent third-party evaluation.

Why does deepening this market matter? Because a narrow market imposes costs in three places. Government borrowing is more expensive and more volatile when there is little competition to set price. Banks locked into sovereign paper lend less to the private sector, reducing investment and growth. And monetary policy transmission weakens when rate changes do not spread across the curve. If 62 percent of banking assets sit in sovereign paper, the banking system is effectively the government's lender, not the private sector's. The document never states that link directly, but concedes it indirectly.

Here is my main disagreement: the plan describes the problem as insufficient market depth, but the problem is really a comfortable dependency between the government and the banks. The first is a technical problem with technical solutions. The second is an incentive problem with political solutions. The document is detailed on the first and nearly silent on the second.

I have seen six months of silence overturn a decision, especially when that decision's cost must be paid before an election. The DMO's semi-annual public reporting is a tool to break that silence. But if the report counts milestones rather than measuring spreads, long-duration demand and private-credit flow, it will be a record of activity, not of progress.

Looking forward, the dates to watch are these. November 2026: formation of the LCBM steering committee. December 2026: publication of the detailed roadmap, a fixed auction release time, and the CDNS review action plan. February 2027: updated DMO staffing and career framework. March 2027: publication of PKRV methodology. June 2027: benchmark policy, liability-management framework, and adoption of the CDNS plan. FY2027/28: revision of the primary-dealer framework to weight quote performance. September 2027: assessment of the securities-lending facility. September 2028: design and launch decision on securities lending, the single-register architecture decision, and major infrastructure reform. The 2028-29 budget: coupon and discount apportionment and collective-investment tax alignment.

Three of these matter most to me. December 2026's roadmap, because it will reveal whether pension and insurance milestones are genuinely included or left vague in Phase III. The FY2027/28 primary-dealer framework, because it will show whether the government truly intends to reward quoting over churning. And the September 2028 single-register decision, because it will show whether Pakistan is centralising its settlement infrastructure or opening it.

One question sits beyond the document but governs everything: is Pakistan prepared to accept a little more uncertainty in its borrowing cost so that a real market can form? If not, the best roadmap will remain only a good document—a fine portrait of a narrow market that never opens its door.