# Briefing — week to Friday, 4 September 2026

*Overheating persists, and the week's sharp jump in crude with a firmer breakeven says price pressure is rebuilding. Corporate spreads widened slightly and index total return fell. Stay short duration, keep credit risk light, hold energy exposure.*

## Regime

The regime classification still reads as overheating, and the model's persistence in that state is high enough that a shift over the coming month is unlikely. That persistence matters more than the classification itself, because it tells the committee that the positioning implied by the regime does not need to be revisited every week. The growth side of the signal is the weaker leg: the gap above trend is slight, and it is the inflation side that carries the case. A classification resting mostly on one leg is worth interrogating rather than accepting, so the question is whether the inflation leg is measured well. The measured inflation gap is negative, which sits awkwardly against a core price series that is running at the top of its recent range and still rising. Two readings of the same economy cannot both be right, and the disagreement is large enough that resolving it decides the positioning. The resolution is timing: the gap is built on a lagged and revised vintage, while the crude move and the firmer breakeven are current, and I weight the current evidence more heavily. A statistic assembled from data that will be revised is a description of the recent past, and the market prices in front of it are a description of now. That argues for keeping duration short and holding energy exposure rather than trimming it. Those two positions are the same trade expressed twice, and I would rather hold both than split the difference.

## Monetary policy

The policy path is asymmetric to the upside, which argues for keeping duration short. Waller framed the next decision as conditional on the August inflation print, saying that if progress continues he is "willing to support holding the policy rate at its current level"<sup>1</sup>, but that "if inflation comes in hot, I would consider a rate hike"<sup>2</sup>. That is a conditional structure with only one active branch: the benign outcome delivers no change, and the adverse outcome delivers tightening. A rate holder therefore earns nothing from the good case and loses in the bad one, which is the definition of an asymmetric distribution around an unchanged policy rate.

The bar for tightening is low, because he judges that "policy is currently only slightly restricting aggregate demand, and it may not take much acceleration in inflation to nudge me into supporting tighter policy"<sup>2</sup>. The size of the required inflation surprise is what determines the probability of the adverse branch, and this framing puts it low. Restraint that is described as slight is restraint that can be argued to be insufficient without any change in the framework, only in the data.

This narrows rather than contradicts the overheating case. Waller sees disinflation underway and does not treat energy as a broad pass-through risk, noting his earlier worry that "higher energy prices would bleed into many goods and services prices hasn't come to pass, at least so far"<sup>3</sup>. The qualification at the end of that sentence carries weight, because it concedes that the absence of pass-through is an observation about the period so far rather than a structural claim. He still flags that "Energy prices have moved up again"<sup>4</sup>. The combination is a policymaker who is not yet alarmed by energy but is watching the same variable the market has just repriced.

A hike is the live risk at this meeting, and short duration is the cheapest way to hold it. It is cheaper than the alternatives because the front end is already priced for no change, so the position costs little if the benign branch arrives.

## Rates and curve

Stay short duration. The move came at the long end and in inflation compensation, not in policy expectations. The ten-year rose +4bp<sup>5</sup> to 4.77%<sup>6</sup>, close to its fifty-two week high of 4.79%<sup>7</sup>, and the thirty-year rose +3bp<sup>8</sup> to 5.25%<sup>9</sup>. The two-year was unchanged at 4.34%<sup>10</sup>, and the upper bound of the federal funds target was unchanged at 3.75%<sup>11</sup>. That combination of a moving long end and a static front end tells you the repricing was not about the timing of the next policy decision.

The decomposition makes the point sharper. The ten-year breakeven rose +4bp<sup>12</sup> to 2.35%<sup>13</sup>, with the real ten-year yield unchanged at 2.42%<sup>14</sup>; the entire nominal move was priced inflation. Nothing in the week's move can be attributed to a higher required real return, to a growth upgrade, or to a change in the term premium demanded for real risk. The two-to-ten spread widened +2bp<sup>15</sup> to 0.41%<sup>16</sup>, a steepening produced entirely by the same mechanism.

A curve steepening driven wholly by breakevens with the front end pinned is the clearest confirmation that price pressure is rebuilding rather than growth. It also means the long end is the wrong place to take interest rate risk, because that is precisely where the inflation compensation is being rewritten. With the ten-year sitting just under the top of its yearly range, the asymmetry in the policy path and the asymmetry in the yield range point the same way.

## Credit

Keep credit risk light. Investment grade spreads widened to 0.81%<sup>17</sup>, up 2bp<sup>18</sup> on the week and 3bp<sup>19</sup> over four weeks, leaving them nearer the fifty-two week low of 0.73%<sup>20</sup> than the high of 0.94%<sup>21</sup>. The corporate total return index fell -0.3%<sup>22</sup> to 3,575.55<sup>23</sup>, a loss of 9.67<sup>24</sup> points. The loss on the total return index is larger than the spread widening alone would produce, which is consistent with the rate move at the long end doing most of the damage to an asset class that carries duration whether or not it is bought for duration.

High yield spreads widened 5bp<sup>25</sup> to 2.65%<sup>26</sup>, though they are still 5bp<sup>27</sup> tighter than four weeks ago and the high yield total return index sits close to its fifty-two week high at 1,914.88<sup>28</sup>. The shorter-duration part of the credit market therefore held up better than the longer one, which again points to rates rather than to deteriorating fundamentals.

The widening is too small to be a credit signal in itself, and the case for carrying less risk rests on duration and inflation rather than on spreads. The problem with these levels is not what they say about defaults but how little they pay for being wrong about rates. I would rather reduce credit because the compensation is thin than wait for spreads to tell me something they are not yet saying.

## Equities

Equities give no reason to add risk here. The S&P closed the week at 7,718.60<sup>29</sup>, up +6.84<sup>30</sup> points, or +0.1%<sup>31</sup>, and is still -39.04<sup>32</sup> points lower over four weeks against a fifty-two week high of 7,798.99<sup>33</sup>. The Nasdaq at 26,584.06<sup>34</sup> gained +0.7%<sup>35</sup> on the week but remains -106.56<sup>36</sup> points below its level four weeks ago. Both indices therefore recovered ground within a month that has gone nowhere, and neither has recovered its high.

Index gains this thin do not offset the case for short duration and light credit risk. A market that stalls while long yields press against the top of their range is not being paid for the discount rate it faces. If the policy risk resolves to a hike, equities are the asset with the least protection in the portfolio, and this week gave no evidence that they are priced for it.

## FX and commodities

Hold the energy exposure. Brent rose to 96.02<sup>37</sup> a barrel, up +6.27<sup>38</sup> on the week and +8.40<sup>39</sup> over four weeks; WTI closed at 91.48<sup>40</sup> after a +6.91<sup>41</sup> gain. The weekly move accounts for most of the four-week move in Brent, so this is a fresh impulse rather than the continuation of a slow drift. The dollar did not amplify the move: the broad index sits at 118.75<sup>42</sup>, down -0.32<sup>43</sup> over four weeks, with the euro at 1.1598<sup>44</sup>.

The oil impulse is therefore domestic in effect, not offset by currency. A stronger dollar would have absorbed part of the crude move before it reached domestic prices, and it has not. This is the current evidence that outweighs the lagged inflation gap, and it is the same evidence the policy branch turns on.

## Macro data

The activity data keep the short-duration stance in place, but the inflation evidence behind it is thinner than the growth evidence. Tracked growth sits at 4.75%<sup>45</sup>, the high of the past year 4.75%<sup>46</sup> and far above the low 1.24%<sup>47</sup>. Payrolls at 159,075<sup>48</sup> are also a one-year high, with unemployment at its low of 4.10%<sup>49</sup>. Growth and labour therefore corroborate each other at the top of their ranges, which is the part of the regime call that does not depend on interpretation.

Core PCE stands at 130.66<sup>50</sup> and headline CPI at 332.81<sup>51</sup>, below its 333.98<sup>52</sup> peak. These are price levels, not proof that inflation are accelerating. Sentiment at 55.20<sup>53</sup> is nearer the year's low than its high.

The asymmetry in the macro pack is that the growth evidence is direct and the inflation evidence has to be inferred, which is exactly why the market-based measures do the work in this call. Weak sentiment alongside strong activity and employment does not change the positioning, because policy responds to the latter.

## Fiscal and trade policy

Policy this week adds to the price-pressure case rather than cutting against it. The beef order directs a review of authorities for mandatory country-of-origin labeling and instructs agencies to weigh the effects, while stating that the national herd is at a multi-decade low and that consumer demand has grown over the past decade: "But the national herd is at a 75-year low"<sup>54</sup>. Labeling and trade restrictions raise landed costs. A supply base described as historically small, facing demand described as having grown, is a market in which added compliance and import friction pass through rather than being absorbed.

Food price pressure is likely to persist into next year, which argues for keeping duration short and holding energy exposure. It matters more than its weight in the basket because it arrives alongside the energy move rather than instead of it, and because supply-side pressure of this kind is slow to reverse. The policy branch that leads to tightening becomes more likely when two separate cost channels are pushing the same way.

## Citations

1. Waller, The Economic Outlook and Some Comments on My Policy Communication ¶19 — https://www.federalreserve.gov/newsevents/speech/waller20260903a.htm (2026-09-03)
2. Waller, The Economic Outlook and Some Comments on My Policy Communication ¶20 — https://www.federalreserve.gov/newsevents/speech/waller20260903a.htm (2026-09-03)
3. Waller, The Economic Outlook and Some Comments on My Policy Communication ¶13 — https://www.federalreserve.gov/newsevents/speech/waller20260903a.htm (2026-09-03)
4. Waller, The Economic Outlook and Some Comments on My Policy Communication ¶17 — https://www.federalreserve.gov/newsevents/speech/waller20260903a.htm (2026-09-03)
5. DGS10.chg_1w = DGS10[2026-09-03] - DGS10[2026-08-28] — https://fred.stlouisfed.org/series/DGS10 (2026-09-04)
6. DGS10.latest = DGS10[2026-09-03] — https://fred.stlouisfed.org/series/DGS10 (2026-09-04)
7. DGS10.hi_52w = max(DGS10[2025-09-08..2026-09-03]) — https://fred.stlouisfed.org/series/DGS10 (2026-09-04)
8. DGS30.chg_1w = DGS30[2026-09-03] - DGS30[2026-08-28] — https://fred.stlouisfed.org/series/DGS30 (2026-09-04)
9. DGS30.latest = DGS30[2026-09-03] — https://fred.stlouisfed.org/series/DGS30 (2026-09-04)
10. DGS2.latest = DGS2[2026-09-03] — https://fred.stlouisfed.org/series/DGS2 (2026-09-04)
11. DFEDTARU.latest = DFEDTARU[2026-09-04] — https://fred.stlouisfed.org/series/DFEDTARU (2026-09-04)
12. T10YIE.chg_1w = T10YIE[2026-09-04] - T10YIE[2026-08-28] — https://fred.stlouisfed.org/series/T10YIE (2026-09-04)
13. T10YIE.latest = T10YIE[2026-09-04] — https://fred.stlouisfed.org/series/T10YIE (2026-09-04)
14. DFII10.latest = DFII10[2026-09-03] — https://fred.stlouisfed.org/series/DFII10 (2026-09-04)
15. T10Y2Y.chg_1w = T10Y2Y[2026-09-04] - T10Y2Y[2026-08-28] — https://fred.stlouisfed.org/series/T10Y2Y (2026-09-04)
16. T10Y2Y.latest = T10Y2Y[2026-09-04] — https://fred.stlouisfed.org/series/T10Y2Y (2026-09-04)
17. BAMLC0A0CM.latest = BAMLC0A0CM[2026-09-03] — https://fred.stlouisfed.org/series/BAMLC0A0CM (2026-09-03)
18. BAMLC0A0CM.chg_1w = BAMLC0A0CM[2026-09-03] - BAMLC0A0CM[2026-08-28] — https://fred.stlouisfed.org/series/BAMLC0A0CM (2026-09-03)
19. BAMLC0A0CM.chg_4w = BAMLC0A0CM[2026-09-03] - BAMLC0A0CM[2026-08-07] — https://fred.stlouisfed.org/series/BAMLC0A0CM (2026-09-03)
20. BAMLC0A0CM.lo_52w = min(BAMLC0A0CM[2025-09-08..2026-09-03]) — https://fred.stlouisfed.org/series/BAMLC0A0CM (2026-09-03)
21. BAMLC0A0CM.hi_52w = max(BAMLC0A0CM[2025-09-08..2026-09-03]) — https://fred.stlouisfed.org/series/BAMLC0A0CM (2026-09-03)
22. BAMLCC0A0CMTRIV.pct_1w = (BAMLCC0A0CMTRIV[2026-09-03] - BAMLCC0A0CMTRIV[2026-08-28]) / BAMLCC0A0CMTRIV[2026-08-28] * 100 — https://fred.stlouisfed.org/series/BAMLCC0A0CMTRIV (2026-09-03)
23. BAMLCC0A0CMTRIV.latest = BAMLCC0A0CMTRIV[2026-09-03] — https://fred.stlouisfed.org/series/BAMLCC0A0CMTRIV (2026-09-03)
24. BAMLCC0A0CMTRIV.chg_1w = BAMLCC0A0CMTRIV[2026-09-03] - BAMLCC0A0CMTRIV[2026-08-28] — https://fred.stlouisfed.org/series/BAMLCC0A0CMTRIV (2026-09-03)
25. BAMLH0A0HYM2.chg_1w = BAMLH0A0HYM2[2026-09-03] - BAMLH0A0HYM2[2026-08-28] — https://fred.stlouisfed.org/series/BAMLH0A0HYM2 (2026-09-03)
26. BAMLH0A0HYM2.latest = BAMLH0A0HYM2[2026-09-03] — https://fred.stlouisfed.org/series/BAMLH0A0HYM2 (2026-09-03)
27. BAMLH0A0HYM2.chg_4w = BAMLH0A0HYM2[2026-09-03] - BAMLH0A0HYM2[2026-08-07] — https://fred.stlouisfed.org/series/BAMLH0A0HYM2 (2026-09-03)
28. BAMLHYH0A0HYM2TRIV.latest = BAMLHYH0A0HYM2TRIV[2026-09-03] — https://fred.stlouisfed.org/series/BAMLHYH0A0HYM2TRIV (2026-09-03)
29. SP500.latest = SP500[2026-09-04] — https://fred.stlouisfed.org/series/SP500 (2026-09-05)
30. SP500.chg_1w = SP500[2026-09-04] - SP500[2026-08-28] — https://fred.stlouisfed.org/series/SP500 (2026-09-05)
31. SP500.pct_1w = (SP500[2026-09-04] - SP500[2026-08-28]) / SP500[2026-08-28] * 100 — https://fred.stlouisfed.org/series/SP500 (2026-09-05)
32. SP500.chg_4w = SP500[2026-09-04] - SP500[2026-08-07] — https://fred.stlouisfed.org/series/SP500 (2026-09-05)
33. SP500.hi_52w = max(SP500[2025-09-08..2026-09-04]) — https://fred.stlouisfed.org/series/SP500 (2026-09-05)
34. NASDAQCOM.latest = NASDAQCOM[2026-09-03] — https://fred.stlouisfed.org/series/NASDAQCOM (2026-09-03)
35. NASDAQCOM.pct_1w = (NASDAQCOM[2026-09-03] - NASDAQCOM[2026-08-28]) / NASDAQCOM[2026-08-28] * 100 — https://fred.stlouisfed.org/series/NASDAQCOM (2026-09-03)
36. NASDAQCOM.chg_4w = NASDAQCOM[2026-09-03] - NASDAQCOM[2026-08-07] — https://fred.stlouisfed.org/series/NASDAQCOM (2026-09-03)
37. DCOILBRENTEU.latest = DCOILBRENTEU[2026-09-01] — https://fred.stlouisfed.org/series/DCOILBRENTEU (2026-09-02)
38. DCOILBRENTEU.chg_1w = DCOILBRENTEU[2026-09-01] - DCOILBRENTEU[2026-08-28] — https://fred.stlouisfed.org/series/DCOILBRENTEU (2026-09-02)
39. DCOILBRENTEU.chg_4w = DCOILBRENTEU[2026-09-01] - DCOILBRENTEU[2026-08-07] — https://fred.stlouisfed.org/series/DCOILBRENTEU (2026-09-02)
40. DCOILWTICO.latest = DCOILWTICO[2026-09-01] — https://fred.stlouisfed.org/series/DCOILWTICO (2026-09-02)
41. DCOILWTICO.chg_1w = DCOILWTICO[2026-09-01] - DCOILWTICO[2026-08-28] — https://fred.stlouisfed.org/series/DCOILWTICO (2026-09-02)
42. DTWEXBGS.latest = DTWEXBGS[2026-08-28] — https://fred.stlouisfed.org/series/DTWEXBGS (2026-08-31)
43. DTWEXBGS.chg_4w = DTWEXBGS[2026-08-28] - DTWEXBGS[2026-08-07] — https://fred.stlouisfed.org/series/DTWEXBGS (2026-08-31)
44. DEXUSEU.latest = DEXUSEU[2026-08-28] — https://fred.stlouisfed.org/series/DEXUSEU (2026-08-31)
45. GDPNOW.latest = GDPNOW[2026-07-01] — https://fred.stlouisfed.org/series/GDPNOW (2026-09-03)
46. GDPNOW.hi_52w = max(GDPNOW[2025-10-01..2026-07-01]) — https://fred.stlouisfed.org/series/GDPNOW (2026-09-03)
47. GDPNOW.lo_52w = min(GDPNOW[2025-10-01..2026-07-01]) — https://fred.stlouisfed.org/series/GDPNOW (2026-09-03)
48. PAYEMS.latest = PAYEMS[2026-08-01] — https://fred.stlouisfed.org/series/PAYEMS (2026-09-04)
49. UNRATE.lo_52w = min(UNRATE[2025-11-01..2026-08-01]) — https://fred.stlouisfed.org/series/UNRATE (2026-09-04)
50. PCEPILFE.latest = PCEPILFE[2026-07-01] — https://fred.stlouisfed.org/series/PCEPILFE (2026-08-26)
51. CPIAUCSL.latest = CPIAUCSL[2026-07-01] — https://fred.stlouisfed.org/series/CPIAUCSL (2026-08-12)
52. CPIAUCSL.hi_52w = max(CPIAUCSL[2025-11-01..2026-07-01]) — https://fred.stlouisfed.org/series/CPIAUCSL (2026-08-12)
53. UMCSENT.latest = UMCSENT[2026-07-01] — https://fred.stlouisfed.org/series/UMCSENT (2026-07-31)
54. Supporting America’s Ranchers ¶1 — https://www.whitehouse.gov/presidential-actions/2026/09/supporting-americas-ranchers/ (2026-09-04)