Why a flexible APR keeps moving

You park some funds in a flexible product, come back the next day, and the headline APR has gone from 5% to 3%. Nothing is broken. Floating is the design. The useful question is narrower: which kind of move was that?

Three completely different things produce the same falling number — your own balance crossed a tier boundary, borrowing demand cooled off, or a limited-time promo ended. They call for different responses, and one of them calls for nothing at all.

Three different kinds of “it moved”

Before you move anything, work out which of these you are looking at:

• A tier boundary. The product pays one rate up to a certain size and a lower rate above it. You did nothing except add funds, and the portion above the line dropped to the next tier. The giveaway: your blended rate fell right after you deposited more.
• Market drift. The rate tracks supply and demand for the asset. You did nothing at all and the number moved on its own. The giveaway: the same product shows different numbers at different times of day or week.
• A promo expiring. A boosted rate ran for a defined window and then reverted. The giveaway: the drop is usually large, and the end date was normally printed somewhere on the page.

Mixing these up leads to a specific wrong conclusion: you triggered a tier change yourself, and you read it as the platform cutting rates across the board. Sorting it out first is what makes the rest of this useful.

Who actually sets the number

A flexible rate sits between two sides. On one side are people who want to lend out idle balances — that is you. On the other are people who want to borrow. When there are many borrowers willing to pay, the lending side has room to move up. When borrowing dries up, there is not much to share out.

That explains something that strikes a lot of people as backwards: rates tend to be higher when the market is busy, not when it is calm. Busy markets mean more people borrowing to do something, which bids up the price of funds. Quiet markets mean the opposite. The APR behaves less like a policy and more like a thermometer.

The second layer is risk pricing. The Chinese Wikipedia entry on credit risk notes that higher risk goes with a higher required rate of interest. Read that backwards and it becomes a practical filter: when a rate is visibly above its peers, the first question is not “what a find” but “what is the extra paying me to accept?” Usually a longer lock-up, a more volatile underlying asset, or a weaker issuer.

Put the two layers together and the daily wobble stops looking mysterious. It is supply, demand and risk pushing on the same number — not a promise being revised.

Tiers: the first slice and the last slice are not priced alike

Tiering is the one that fools people. The usual shape is a higher rate up to some size, and a lower rate on everything above it. The boundary, the number of tiers and the rate in each one differ by product and change over time — read the product page you are actually looking at. This article gives no specific figures, because any figure printed here would go stale.

Here is why it misleads, with a deliberately hypothetical example: suppose a product pays 6% on the first 1,000 units and 2% above that. Deposit 1,000 and your blended rate is 6%. Add another 1,000 and the blended rate falls to roughly 4% — you are not earning less in absolute terms, you are earning more, but the headline number on the page genuinely went down.

Those numbers illustrate a structure and describe no real product. The point to keep: a falling blended rate after a deposit is not automatically bad, but you should know what the marginal slice is actually earning. If the above-tier rate is poor enough that you would not have chosen it on its own, that money probably belongs somewhere else rather than stacked on top.

What the word “estimated” is doing there

Rate displays usually carry a qualifier: “estimated”, “reference”, “last 7 days”. Those are not politeness. They tell you where the number came from. “Last 7 days” looks backwards at what actually happened; “estimated” projects forward from the current state. Neither is a commitment about the future.

What determines your actual return is the real rate on each day the funds were earning, not the figure on screen at the moment you deposited. In a period when rates move a lot, that gap gets wide fast, and any “expected return” you calculate by multiplying the screen number by 365 is a rough sketch at best.

So read the number as a snapshot of now, not a rate contract. If you want an actual estimate, run a few scenarios at different rates in the earnings calculator rather than fixating on one figure.

When it is worth acting, and when it is not

Match each kind of move to a response:

• Dropped into a lower tier. Worth doing the arithmetic. If the above-tier rate is clearly worse than your alternatives, moving that slice out is reasonable. Leave the part still earning the top tier where it is.
• Small market drift. Usually not worth touching. Moving funds has friction — redemption waits, a fresh accrual start, and your own attention. Chasing a point or two of short-term drift back and forth rarely nets out ahead.
• Promo ended. Worth re-evaluating. You went in for the boosted rate; the boost is gone, so the reason to stay is gone with it. Compare it against alternatives again at the ordinary rate.

One more case that argues for doing nothing: seeing a higher rate elsewhere and moving without checking why it is higher. As above, the extra is usually compensating for something. Before moving, at least name the something and decide whether you want it.

And if you have not yet settled whether this money can sit still at all, that question comes before rate-shopping — it is the subject of how to pick a product.

Common questions

I did not touch anything — why did my rate drop?

Usually one of two things. Either your balance crossed a tier boundary, so the portion above the line now earns a lower rate and drags the blended figure down; or borrowing demand fell and the floating rate followed it. The first is triggered by your own deposit; the second has nothing to do with you. Current tiers and rates are whatever the product page shows at the time.

A rate is far above everything comparable. Should I take it?

Start by naming what the extra is compensating for. Common answers: a longer lock-up, a more volatile underlying asset, or a weaker issuer. That does not automatically make it a bad choice, but it does mean you are being paid to accept something. Identify it before deciding, instead of comparing headline numbers.

Can I treat the “estimated APR” as what I will actually receive?

No. Words like “estimated”, “reference” or “last 7 days” mean the figure is either backward-looking or projected from current conditions — neither is a promise. Your actual return depends on the real rate on each accruing day. Model a few rate scenarios instead of multiplying one screen figure by a number of days.