Job-hopping is
just repricing.

Run the math and the “unfair” result turns out to be arithmetic.

What the numbers actually say

The average Malaysian increment for staying put is around 5% a year — employers themselves forecast 5.33% for executives in 2026. Compound 5% for fifteen years and pay roughly doubles. Start anywhere near a normal base and you land almost exactly at RM8.5K.

Switching jobs reprices you at 15–20% per move — and that is the floor, not the ceiling; almost half of Malaysian job-changers got more than 20%. Three hops at 20% is a 73% jump. That is your RM12K junior.

Nobody cheated. The senior let one employer price him for fifteen years. The junior went to market three times and let the market do it. The increment system pays loyalty in percentages of a stale base. The job market pays in current value. Those two numbers were always going to diverge — the viral post just put faces on them.

Fifteen years of what?

Here is the harder question hiding under the salary one: what exactly was the senior selling for fifteen years?

Years of experience is an input. Nobody buys inputs. A customer pays for output — the thing produced, shipped, solved, closed. If fifteen years of experience shows up in the output, it gets paid for. If it shows up only on the résumé, it is decoration. I have made this point before: experience only counts where it touches reality.

AI makes this brutal, because AI compresses exactly the layer that tenure used to guard: the routine competence, the “I know how we do things here,” the procedural memory that took a decade to accumulate and now ships inside a model. The reset puts everyone back on the same starting line — and what is left to pay for is output and judgment.

Do the uncomfortable audit on those fifteen years. How much of it is knowledge a model now carries — the formats, the procedures, the “standard way we handle this”? Whatever part of your experience can be prompted is no longer a salary premium. It is a subscription, rentable by anyone for less than a lunch. The premium that survives is the part no model ships: judgment under consequences, and the ability to be answerable for a result. A pay system indexed to years is pricing the asset that is depreciating fastest.

It cuts the other way too. AI breaks the link between time served and output produced. I have watched an AI-capable beginner out-produce people with ten years on the clock — not because the beginner is smarter, but because the leverage moved. When time stops predicting output, every pay system indexed to time is pricing noise: the seniority scale prices the wrong fifteen years, and the hopper’s résumé prices the wrong three moves.

And AI ends the hiding, on both sides. When output is generated, tracked and reviewed through the same tools, it becomes measurable per person, per week. The fog that let tenure pass as output is lifting — and so is the fog that let a polished résumé survive probation. The measurement the senior avoided and the hopper escaped is becoming automatic. Pay was always going to meet the market. Now it meets the meter.

How I pay at IMA

I will put my own payroll where my mouth is. At IMA, the rate is set by output — not by year of experience. And our people are paid well above market average. Those two facts are not a coincidence; the second is what the first produces.

When someone’s output doubles, the pay follows. I do not check the calendar first. I do not check what year they graduated. The question is never “how long have you been here” — it is “what does the business collect because you are here,” and the answer is shared with the person who produced it.

Why would I run it any other way? My best people are exactly the ones the market reprices most eagerly. If I pay them a stale base plus 5%, the market will happily correct my mistake — at 20% per correction, on someone else’s payroll. Paying for output is not generosity. It is keeping the asset.

The squeeze model is a countdown

Whenever I see this kind of post, my answer is the same, and it never changes: businesses that 压榨员工 — that squeeze their people and bank the gap — will not be reformed. They will be disqualified.

Underpaying is a business model. Some bosses’ entire margin is the space between what their people produce and what they are paid. That model worked when employees had no price discovery — when you did not know your own market value, and moving was rare and shameful. Every one of those conditions is gone. Salaries are one viral post away from public. Moving is normalised at 20% a hop. Loyalty stopped being free.

And a new generation of bosses is arriving that treats sharing the result as the operating system, not a perk — because they know the arithmetic above better than the incumbents do. The squeeze-model boss will not lose an argument. He will lose his people, one reprice at a time, to businesses built to share. It is the same expiry I keep writing about: time does not disqualify people — it disqualifies models.

Don’t read me as pro-hopping

Now let me disappoint the other half of the comment section: I do not see job-hopping as good practice either. Repricing is real — but a hop proves nothing about capability. A market offer prices your résumé. Only output prices you.

There is one hop I fully accept: the first one. A fresh graduate has no experience, so the first job often does not match what he imagined. Leaving that mismatch is learning. But hopping too often is its own confession — it says you do not have the capability to understand what you actually want. You are not repricing; you are wandering, at 20% a move.

And I will add something from my own hiring experience, because the pattern is too consistent to be polite about: many frequent hoppers move because they cannot deliver. Watch the timeline, not the story. The stays that end right as probation ends — that is not ambition, that is exit before discovery. They leave before the company finds out their real capability, then sell the clean résumé to the next employer, who pays 20% more for the privilege of running the same probation.

So the RM12K only stands if output eventually backs it. The market reprices upward fast — and corrects downward just as fast, the day someone finally measures. Tenure without output and hops without output are the same disease wearing different clothes: both are asking to be paid for time instead of results.

To be fair about both sides

Not every long stay is stagnation — some seniors are underpriced precisely because they compound quietly where nobody benchmarks them. And output-based pay is harder to run than a pay scale. You must actually measure output, argue about it honestly, and pay up when the number is good — including when it embarrasses the org chart.

Plenty of businesses also genuinely cannot pay more, because their model does not produce enough per head to share. But notice what that means: the model cannot afford the market. That is not the employee’s problem to subsidise — and increasingly, they won’t.

The bottom line

For employees: get repriced on purpose. Once a year, bring one of two things — a market offer, or an output case your boss cannot argue with. Loyalty is fine; unpriced loyalty is a donation.

For bosses: index pay to output and share the result before the market shares it for you. The RM8.5K post is not a story about one underpaid man. It is the market publishing your countdown, one viral share at a time.