A deposit paid above the insured rate will leave first

Every mid-sized bank in Indonesia knows the telephone call. The treasurer of a company, a foundation or a wealthy family rings a branch near the end of the month, names a balance and asks what rate we will pay. If our answer is below what another bank offered that morning, the money moves. Special-rate deposits are a normal part of our market and I do not think they should be prohibited. My argument is narrower. A deposit priced above the rate that LPS uses to decide whether it will insure a deposit should be treated by management and the board as market funding: capped, disclosed and priced as though it could leave within a week, because it can.

What the insured rate tells a depositor

Indonesia's deposit insurance, run by LPS, covers each depositor at each bank up to Rp 2 billion, but only where the interest paid does not exceed a rate that LPS sets and revises periodically. A depositor who accepts a special rate above that level has made a choice, whether or not they would describe it so. They have exchanged insurance for yield. Most of the time the exchange costs them nothing. In a period of stress, when a rumour about a bank travels through chat groups faster than the bank can answer it, the uninsured depositor has every reason to leave first and ask questions later.

I learnt to respect that logic in 1998, watching depositors abandon Indonesian banks faster than any model had assumed. I relearnt it more quietly in 2022, when rates rose and our special-rate balances repriced within weeks while our savings accounts barely moved. At the end of 2021, special-rate time deposits made up about a fifth of Kartadana's third-party funds. They were concentrated among a few hundred depositors, and the average time a balance stayed before it was renegotiated or withdrawn was under four months.

Those numbers did not alarm our treasury team, and they were not wrong to be calm. The balances had rolled over reliably for years. My concern was what that reliability rested on. It rested on our paying the market rate every time, which meant our funding cost was set by our most price-sensitive customers, and on those customers never having a reason to doubt us. A bank that depends on both conditions has a funding plan for good years only.

Pay for transactions, not for balances

The alternative is unglamorous. It is to earn deposits that stay because they are used. For a bank that lends to small businesses, that means the current accounts of the businesses we already finance and the savings accounts of the people those businesses pay. Since 2022 we have put most of our deposit effort into three things: QRIS merchant acquiring for traders and shops, so that their daily takings land in a Kartadana account; payroll services for workers in the market and manufacturing clusters where we lend; and supplier-payment tools for distributors, so that the small businesses at the end of a supply chain are paid into accounts with us.

None of these produces a large balance quickly. A market trader's account holds a few million rupiah at any moment. But the balances are numerous, they move for reasons that have nothing to do with our rate, and each one tells us something about a business we may lend to. A loan officer who can see six months of a trader's QRIS receipts knows more about that trader than a set of annual accounts would show. The deposit and the credit decision reinforce each other, which special-rate funding never does.

Between 2021 and June 2026, current and savings accounts rose from 31% to 46% of our third-party funds, and special-rate time deposits fell by Rp 6.4 trillion. We now cap special-rate balances at 12% of third-party funds, report them to the Board of Commissioners separately from ordinary time deposits and stress them on the assumption that most could leave within a month. I would not claim that the cap suits every bank. It is the level at which our liquidity buffers could absorb a sudden departure without forcing us to sell assets or to bid even higher for replacement money.

The price of a slower balance sheet

The case against my position is real, and our finance director made it well. Special-rate deposits are the quickest way for a mid-sized bank to fund loan growth. Turning them away in 2022 and 2023 meant growing loans more slowly than we could have, just as demand from small businesses was recovering. Transaction banking requires investment in merchant devices, payroll software and people before it yields a single rupiah of deposits. For about two years our deposit growth trailed the industry, and an analyst asked at a results briefing whether we had lost our competitive position.

We also lost relationships. Several large depositors moved their money to banks that would match any rate, and some took other business with them. I accepted that, because a relationship that exists only at the top of the rate table is not one we could rely on when it mattered.

Against those costs, our cost of funds is roughly where it was in 2021 even though policy rates were higher for much of the period, our liquidity measures carry more headroom and the deposits we gained are tied to the borrowers whose credit we most need to understand. The benefit I value most is harder to express as a ratio. When a rumour about mid-sized banks circulated in our markets in 2025, our deposit base barely moved, because few of our depositors had any reason to be first out of the door.

If I were advising the board of another Indonesian bank, I would ask for a single table: deposits split by whether they are priced within or above the insured rate, with the average time each category stays and the names of the twenty largest special-rate depositors. I would then ask management what the loan book would look like if half of that category left in a month. The answer will not always justify a cap. It will tell the board what kind of funding plan it has actually approved.