Japan Wage Data: Reading the Monthly Labour Survey and the 3% Wage Benchmark

Japan’s most market-sensitive wage number is not the headline. The headline of the Monthly Labour Survey — total cash earnings — is dominated by bonus timing and sample noise; the series analysts watch most closely for Bank of Japan policy signals is scheduled (base) pay for full-time workers on a matched sample of establishments, and sustained growth in the neighborhood of 3% there is a common analyst benchmark for when further rate hikes become likelier. The Bank itself weighs a broad set of wage and inflation indicators rather than any single series or published threshold, but if you learn to skip the headline and go straight to scheduled pay, you will read the release better than most of the fast-money reaction.

The stakes are live: with the yen trading around 160 per dollar in late August 2026 and JGB yields pricing continued tightening, each monthly wage print feeds directly into the rate debate. But wage data is a recurring monthly event, not a one-off. What follows is the durable part: how the release works, every month, in any year.

What the Monthly Labour Survey Is and When It Lands

The Monthly Labour Survey (Maitsuki Kinrō Tōkei Chōsa, often shortened to “Maikin”) is compiled by the Ministry of Health, Labour and Welfare (MHLW). It surveys establishments with five or more employees and reports average monthly earnings, hours worked, and employment, split by full-time and part-time status.

Two calendar facts matter for traders:

Timing. The preliminary estimate for a given month is typically published roughly five weeks later — which generally places it in the first half of the second following calendar month — at around 8:30 a.m. Tokyo time, before the cash equity open, so the rates and FX reaction comes first.
Two bites. A revised estimate follows a few weeks after the preliminary one. Revisions are not cosmetic: they routinely shift year-on-year growth by several tenths of a percentage point and have occasionally flipped the sign of real wage growth. Put both dates in your calendar, not just the first.

Three Wage Numbers in One Release

Every release contains at least three distinct wage concepts. Confusing them is the single most common reading error.

Series	What it includes	How to use it
Total cash earnings (nominal)	Base pay + overtime + bonuses/special payments	The headline. Volatile; dominated by bonus timing in summer and December.
Real wages	Total cash earnings deflated by CPI (historically the “all items less imputed rent” series)	The political and household-purchasing-power number. Mechanically lags nominal wages when inflation runs hot.
*Scheduled earnings (shoteinai kyūyo)*	Contracted base pay only — no overtime, no bonuses	The cleanest read on underlying wage trend and on shunto pass-through. This is the series analysts cite most often when handicapping BOJ policy.

Two supporting series round out the picture. Overtime pay (shoteigai kyūyo) is a cyclical demand signal — firms adjust overtime before headcount. Special payments (bonuses) spike in the June–August and December prints; a big headline surprise in those months is usually a bonus-timing artifact, not a trend change.

Why the Headline Is Noisy: Samples, Bonuses, and Revisions

The Monthly Labour Survey has well-documented statistical quirks, and the market has been burned by each of them:

Sample rotation. Part of the establishment sample is replaced periodically. When the sample changes, measured average wages can jump or drop for purely compositional reasons. The fix is the “common establishments” (kyōtsū jigyōsho) series, which computes year-on-year growth only across establishments present in both periods. When the full-sample and common-establishment growth rates diverge, trust the common-establishment figure.
Composition effects. The headline average blends full-time and part-time workers. A rising part-time share drags the average down even when everyone’s hourly pay is rising. Read the full-time and part-time breakdowns separately.
A history of measurement problems. The survey went through a public sampling-irregularities scandal in the late 2010s that forced retroactive corrections. The methodology has been reinforced since, but the episode is why seasoned readers treat any single month’s print as an estimate, not a fact.
Preliminary-to-revised drift. Large bonus-paying firms report late, so bonus-heavy months are especially prone to revision. Never build a narrative on a preliminary summer print.

The Shunto Cycle: How Spring Negotiations Reach the Monthly Data

Shunto — the “spring wage offensive” — is Japan’s annual synchronized wage negotiation between enterprise unions and management, coordinated nationally by Rengo (the Japanese Trade Union Confederation). Rengo publishes running tallies of agreed wage increases from roughly March through early summer. Recent rounds have produced the strongest headline settlements in roughly three decades.

Three mechanics determine how shunto shows up in the Monthly Labour Survey:

Decompose the headline. A shunto settlement combines base-up (bēsu appu — an increase in the base wage scale itself) and teiki shōkyū (automatic seniority-based increments, historically worth roughly 1.5–2 percentage points). Only the base-up portion lifts the economy-wide average wage; seniority increments largely cancel out in aggregate as older workers retire and younger ones join. A “5% shunto” is therefore closer to a 3% aggregate wage impulse.
Expect a lag. Agreements are struck in spring but implemented from the new fiscal year (April), and they filter into paychecks — and into the survey’s scheduled-earnings series — over April through summer. The April-to-July prints are where you verify whether the negotiated numbers are actually reaching the data.
Mind the coverage gap. Shunto tallies over-represent large unionized firms; the survey covers small establishments where wage growth is usually slower. Scheduled-earnings growth in the survey will normally print below the Rengo base-up tally. The question to ask each spring is whether the gap is narrowing — that is the sign of wage growth broadening beyond large caps.

The 3% Benchmark: An Analyst Heuristic, Not an Official Trigger

The Bank of Japan frames its mandate as achieving 2% inflation sustainably, underpinned by a “virtuous cycle” between wages and prices. The commonly cited 3% wage threshold is an analyst heuristic derived from that framing, not a number the Bank has published as an official policy trigger. The arithmetic behind it is simple: roughly 2% inflation plus roughly 1% trend productivity growth implies that nominal wages need to grow around 3% for the inflation target to hold without squeezing real incomes.

In practice, the Bank evaluates a broad set of wage and inflation indicators rather than any single series — scheduled full-time pay on the common-establishment basis is one prominent input, cross-checked against shunto tallies and services inflation (which is where labor costs pass through to prices), among others. One strong month does not settle the question; the operative word in BOJ communication is sustained. That is why a single upside surprise rarely triggers a policy shift by itself, while a string of prints holding near 3% shifts the entire rate-hike probability path.

Worked Example: Reading a Wage Print Against the Rates Market

A wage release means little in isolation — it matters relative to what the bond market already prices. Here is how to do that read with real numbers. As of the August 27, 2026 close (Ministry of Finance official yields), the 2-year JGB yielded 1.696% and the 10-year 2.897%.

Start with the 2-year. The 2-year yield approximates the average policy rate the market expects over the next two years. At 1.696%, it embeds a path of continued hikes from the prevailing policy setting — the market is already leaning toward further tightening.
Score the print against that path. If the next Monthly Labour Survey shows scheduled full-time pay (common establishments) holding around 3%, it validates the priced path — which could nudge the 2-year up modestly, though the actual reaction depends on what was already expected, on positioning, and on whatever else hits the tape that morning. For context on how rate-sensitive equities have been trading, the Banks sector — a classic beneficiary of higher rates — had gained 4.1% over the five sessions to August 27, 2026 (TOPIX-17 ETF proxy).
Diagnose before trading a miss. If the headline misses badly, check which component moved. A miss driven by special payments in a bonus month, with scheduled pay intact, should not reprice the 2-year once analysts strip out the noise. If the 2-year drops sharply anyway, the move may be tracking the headline rather than the underlying trend — a diagnosis worth re-testing against the revised release, with no guarantee about how or when the market reprices.

The discipline: never react to the headline delta versus consensus. React to the scheduled-pay trend versus the 2-year yield.

A Reading Checklist for Release Morning

Confirm whether you are reading the preliminary or revised figure.
Skip total cash earnings; pull scheduled earnings, full-time, common establishments.
Check the real-wage deflator: when food and energy inflation run hot, the “less imputed rent” CPI deflates wages harder, and negative real wages can coexist with a healthy nominal trend.
In the June–August and December prints, discount special payments before drawing conclusions.
From April through summer, compare scheduled-pay growth with the latest Rengo base-up tally — pass-through is the story.
Cross-check the 2-year JGB reaction against your component diagnosis before assuming the market read it correctly.

FAQ

When exactly does Japan’s wage data come out?

The MHLW publishes the preliminary Monthly Labour Survey roughly five weeks after the reference month — typically in the first half of the second following calendar month, at around 8:30 a.m. Tokyo time — with a revised estimate a few weeks later. Exact dates shift month to month, so check the MHLW release calendar rather than assuming a fixed day.

Why can real wages be negative while nominal wages are rising?

Real wages are nominal earnings deflated by consumer prices (historically the CPI “all items less imputed rent” series, which runs hotter than core measures when food and energy climb). Whenever inflation exceeds nominal wage growth, real wages print negative — a statement about purchasing power, not about whether the wage trend is improving.

Do strong shunto results guarantee strong monthly wage data?

No. Shunto tallies cover large unionized firms and include seniority increments that do not lift the aggregate average; the survey covers small establishments too. Expect survey-based scheduled-pay growth to run below the shunto headline, with the negotiated gains appearing gradually from April onward. Watch whether the gap narrows.

Which series matters most for BOJ policy?

Analysts most often point to scheduled (base) pay for full-time workers on the common-establishment basis, sustained — not one month — around 3%, alongside evidence that labor costs are passing into services prices. But the 3% figure is an analyst rule of thumb rather than an official BOJ threshold, and the Bank itself evaluates a broad range of wage and inflation indicators, including shunto outcomes and services prices, in describing the wage-price virtuous cycle.

Sources

Disclaimer: This is an information and analysis publication, not investment advice. See our Methodology for data sources, standards, and our corrections policy.