9.64% Isn't Everyone's Report Card: Alongside GDP, Taiwan Is Still Missing a Distribution Table

GDP measures the value of final goods and services produced domestically over a period. It can tell us whether the overall size of the economy is expanding and which sectors are contributing to output; it also helps government estimate tax revenue and gauge the business cycle. Without GDP, public debate would too easily be reduced to pure feeling.

But GDP is not a paycheck, and it is not a household balance sheet. A unit of new output might become wages, corporate profit, depreciation, tax revenue, or overseas investment income; it does not automatically tell us which category of employees worked more hours, which region is under rent pressure, or whether non-tech firms are winning orders.

Most importantly, several doors stand between GDP's "real growth" and a household's "disposable slack." Has a company converted its added profit into wages and domestic investment? Has inflation eaten into nominal income? Is a household holding back on spending because of a mortgage, childcare or long-term care costs? None of these are questions a single GDP figure can answer for us.

The Ministry of Finance reported that exports grew 47.1% year-on-year in the first half of 2026. ^[2] That figure shows external demand is extremely strong, and that Taiwan's high growth is now tightly linked to the global tech investment cycle.

But an export surge can't be directly translated into an across-the-board income gain. Exporters, suppliers, logistics, services and the local labor market are hit by this transmission at different speeds; and the output created by a capital-intensive industry is not distributed evenly across all workers by headcount.

This is not meant to paint the tech sector as a problem. On the contrary, AI demand is Taiwan's most valuable updraft right now. The issue is that when an updraft is this concentrated, policy cannot just ask how strong it is — it also has to ask whether it can spread: are non-tech industries getting equipment, talent and orders? Do local governments have the capacity to absorb the investment? Is corporate profit being converted into R&D, training, and more stable jobs?

If only the aggregate is published, concentrated growth gets repackaged as everyone benefiting at once; if output and diffusion are reported separately, government actually knows where the next policy bottleneck sits.

DGBAS reported that in May 2026, the average regular earnings of full-time domestic employees was NT$52,164, and the median was NT$41,050. ^[3]

Both numbers are real, but they are not answers to the same question.

The average sums all wages and divides by the number of employees, so it is easily pulled upward by high earners; the median ranks employees and takes the middle position, which is closer to a typical employee. But the median is not a complete picture of lived experience either: it does not capture all self-employed workers, asset income, or differences in household size, and it has not yet been adjusted for rent, mortgage payments, childcare, long-term care, or energy costs.

Mature economic communication should therefore not pick whichever of the average or the median is most flattering. Both should be published, together with a clear statement of the population, the time period, whether the figure is inflation-adjusted, and how it differs from household disposable income.

The CPI rose 2.60% year-on-year in June, with a first-half average increase of 1.70%. ^[4] That is an aggregate price indicator, not proof that every household faces the same basket of costs; households with a mortgage, renters, families raising children, and energy-intensive businesses face very different price pressure to begin with.

"Not feeling it" is not an anti-scientific mood — it is a reasonable reaction from a different position in the distribution to the same aggregate number.

Those opposed to adding more indicators have a legitimate worry: economic policy is already complicated enough; adding tables for wages, regions, industries and spending could leave decision-makers staring at charts instead of acting.

That worry is valid, which is exactly why a second table cannot become an endlessly expanding wall of data. It has to serve a clear decision, and every set of numbers in it has to answer: "given what we see, what do we do differently?"

For example, if orders and employment in non-tech industries keep weakening, the policy response should not simply be announcing another aggregate growth rate — it should examine bottlenecks in financing, energy, talent and export markets. If median real wages stagnate while the average rises, public communication should explain which sectors the growth is concentrated in, rather than implying "everyone is getting richer."

The second table is not meant to replace GDP — it is meant to stop GDP from being asked to answer questions it was never built to answer.

This table does not need to be updated weekly, but it does need to be fixed, comparable, and something people can be held accountable to. It should have at least four cells:

Typical employees: median real wages, working hours, and the distribution across low, middle and high earners. This answers whether growth is reaching workers.

Industry and region: output, employment, investment and exits for tech versus non-tech industries, broken down by major region. This answers whether the updraft is leaving a handful of sectors.

Essential household spending: housing, food, transport, childcare, long-term care and energy. This answers how much choice a household has left after its income rises.

Diffusion channels: whether corporate profit is being converted into domestic R&D, local procurement, talent development, public tax revenue, and second-tier industrial capacity. This answers whether high growth today can become tomorrow's shared foundation.

DGBAS is responsible for the statistical methodology, the Executive Yuan for cross-ministry response, and local governments and industry intermediaries for explaining their own numbers. Only by placing responsibility right alongside the table can this second scorecard avoid becoming just another piece of propaganda.

If a table is only published when the numbers look good, it is still a propaganda tool. The second scorecard must use the same methodology whether the cycle is going up, going down, or being revised, so readers can track whether "who benefits, who is under pressure" is actually changing.

It also cannot just report the outcome. Every cell needs three additional columns: when the data was last updated, the direction of change from the prior period, and which policy tool is responsible for responding. When average wages rise but median real wages fall, readers should see the distribution problem; when non-tech output holds steady but employment drops, readers should see the tension between productivity and jobs; when essential household spending rises, readers should see the cost implications for subsidy, housing and care policy.

Doing this makes public communication harder, but it makes policy more testable. Ministries can't just cherry-pick the best number, and the media doesn't get to overturn an entire set of national accounts with a single feeling. When an indicator turns worse, government has to explain whether it's a short-term base effect, an industry shift, or policy that hasn't kept up; when an indicator improves, it also has to say whether that improvement is spreading to more people.

What the second table is really meant to build is not a vote on "does everyone feel good," but a chain of accountability people can trace back through.

For the state, the question is how to turn the AI upswing into a thicker power grid, education, care system and non-tech productivity — rather than treating a one-off high-growth year as permanent tax revenue.

For industry intermediaries, the question is which suppliers, talent pools and regions haven't connected to the orders, and how to turn diffusion indicators into concrete procurement, training and financing partnerships.

For SMEs, the question is not "did GDP hit 9.64%," but whether their own orders, margins, wages and cash flow are keeping pace. Aggregate growth can only provide context — it cannot substitute for a firm's own read on demand.

All three kinds of decisions need GDP, but none of them can rely on GDP alone.

Taiwan does not need to apologize for 9.64%. Catching AI demand in the middle of global war, energy and trade friction is the payoff of years of industrial accumulation.

But for good news to become public trust, it has to survive a second question: has this growth, beyond making Taiwan run faster, left more people standing on steadier ground next year?

The point of a second scorecard is not to dismantle GDP, but to finish telling the story of who is accountable for growth. The aggregate is real, and the distribution is real too; only by putting both on the same table does Taiwan earn the right to ask what the next 9.64% should buy.