Nothing got worse. One of three numbers stopped getting better, and the economy's label changed.

Where we are

The market-risk light reads YELLOW, as it has on every daily read since September 7. It is not near either of its lines this week. The light turns RED only when consumer sentiment is below 55 and the world's central-bank money is draining at the same time. Sentiment is 51.7, from the University of Michigan survey of August 28, so the first condition is met; the money gauge reads 'supportive', so the second is not. GREEN would need sentiment at 70 or above. The next sentiment reading is due September 25.

The change is in the second read, the one that describes where the economy is. Jawz places it in one of four states from two directions: is growth rising, and are prices rising? For the last four weeks both were rising, which Jawz calls 'economy running hot' (label SUMMER). From today, prices are still rising but growth is no longer rising — it is flat. Flat growth with rising prices is 'growth stalling, prices rising' (label FALL).

Growth is scored from three measures: factory output, new claims for unemployment benefits, and consumer sentiment. From September 9 to September 18 all three read as improving. On September 19 the factory measure went to flat. Claims and sentiment are still improving, so the overall growth score is still the best of its three grades. The direction, which is what the four states use, went from up to flat. Nothing in the data got worse. One measure stopped getting better.

The label did not change on September 19. Jawz only moves it after three daily reads in a row agree, so September 19 and 20 were reported as a pending change, and today's read is the third. That is also why the weekly history tool still shows September 20 as 'running hot': on that day the change was two days old and not yet confirmed.

Honest caveat: this is a line the label keeps crossing. It moved from 'running hot' to 'growth stalling' in mid-July, back to 'running hot' in late August, and across again today — three crossings in twelve weeks, and today's growth direction sits exactly at flat. The tool marks today's read as on the boundary for exactly that reason. Inflation is still the dominant risk.

China's August statement arrived, and the yen cancelled it

Last week's edition said China's central bank would publish its August statement between September 15 and October 5 and that this brief would report whatever it showed. It came out on September 15. China's central bank shrank its balance sheet in August, from ¥50.207 trillion to ¥49.857 trillion — down 0.70%, the first fall after rises in June and July. In dollars, China's part of the four-bank total went from $7.48 trillion to $7.43 trillion.

On the same day, the total did not move: $25.22 trillion before and after. The yen strengthened from 156.11 to 153.71 per dollar, and the Bank of Japan's part rose from $4.13 trillion to $4.19 trillion. The Bank of Japan figure is monthly and the latest on file is still August's, so that rise is the currency, not new money. A fall in China and a move in the yen landed on the same day and roughly cancelled. The weekly history tool dates the China step to September 20, which is its sample date; the daily record says September 15.

The money gauge still reads 'supportive', on every daily read this week. The trend this brief trusts most — the Fed, the European Central Bank and the Bank of Japan, measured the same way at both ends — went from $17.81 trillion on July 2 to $17.80 trillion now: flat, down 0.06% over twelve weeks. Last week the same measure was down 0.62%; most of that difference is the window's start date moving forward a week, plus this week's yen. China's statement is now 50 days old, down from 75.

Borrowing got more expensive again, for a different reason

The real cost of borrowing for ten years — the ten-year US government bond yield minus the inflation the market expects — rose from 2.55% to 2.61%, again the highest of the thirteen weekly readings on file. It was 2.25% on July 2. This is the sixth edition in a row naming it as the one measure that keeps moving the same way.

Last week it rose because the borrowing cost itself rose. This week the borrowing cost barely moved (4.95% to 4.94%); what changed is that the market's inflation forecast fell, from 2.40% to 2.33%. Same headline number, opposite cause. It is also worth setting beside the economy label: the label says prices are rising, and it is built from monthly price reports. The market's own daily forecast of inflation eased this week. One week of a daily number does not overturn two monthly trends, but the two are not saying the same thing right now.

The rest was calm. The overall direction of financial conditions reads 'tightening' again. Riskier companies still pay 2.70% extra to borrow; safer ones pay 0.78%. Expected stock-market volatility fell from 17.8 to 15.4, calm. The dollar is down 2.21% over twelve weeks.

What to watch

• The factory measure. It is the one that moved the label. If it turns back to improving, the label crosses back after three days; if a second growth measure goes flat or worse, the label stops sitting on the line.
• The consumer survey on September 25. It is one of the three growth measures and the first switch of the warning light.
• The real cost of borrowing at 2.61%, and which half is moving it — the borrowing cost or the inflation forecast.
• Data quality is clean: every source current, nothing stale or unavailable, the collector healthy, and every weekly row behind this edition recorded live. Two instrument issues stay open and are repeated here until they close: the weekly regime history still shows no RED since August 3, although the daily record holds six RED days from September 1 to 6; and the August 17 conditions-direction discrepancy is unchanged for a sixth week. The September 3 edition's date error, listed here last week, was corrected on September 14.

The frame, not the trade

The way to be wrong this week is to read 'growth stalling' as news that the economy weakened. It did not, on these numbers. A summary built from directions can change without anything getting worse — it only needs something to stop getting better. And a summary that has crossed the same line three times in twelve weeks is telling you more about where the line is than about where the economy is going.

The general lesson: before acting on a label that changed, find the smallest change in its inputs that would have produced it, and check whether something that small just happened. Here the answer is one of three measures moving one step, from improving to flat.

So the question for your own book: is any holding there because of a label — a sector call, a 'late-cycle' view, a theme — rather than the numbers under it? If that label changed tomorrow for a reason this small, would you act on it? If yes, the position rests on the label, and you might want to write down which of the underlying numbers would actually change your mind.

None of this is advice and none of it is a forecast. It is a week's data, the reasoning applied to it, and a changed label traced back to the one number that moved it.