#美国8月PPI录得5.4%高于预期

The August PPI print is out and it brought no relief to anyone hoping this inflation story was over. Headline producer prices rose 0.4% on the month, exactly in line with consensus, but the year over year number came in at 5.4%, a tenth above the 5.3% economists were looking for and clearly above July's roughly 4.7% to 4.8% pace. That is the strongest jump in wholesale inflation in three months. Strip out food and energy and core PPI rose 0.2% on the month and about 4.6% on the year, up from something like 4.2% in July. That is roughly two and a half to three times the Fed's 2% target. The composition matters: this is an energy and food story, not a broad demand story. Brent crude pushed above 100 dollars a barrel to 101.21, West Texas Intermediate settled near 96.05, gasoline is back above 4 dollars a gallon and diesel printed a record 5.94 dollars a gallon. A supply shock coming out of a war in the Middle East does not respond well to higher interest rates, which is exactly what makes this so uncomfortable for the Fed.

Wall Street is not waiting for permission to reprice the rate path. Fed funds futures now imply roughly a 60% probability of a 25 basis point hike at the September 15 and 16 FOMC meeting, a genuinely unusual place to be, and the fact that a majority of economists in the latest Reuters poll still expect a hold is exactly why the market is jumpy. The ten year Treasury yield has climbed to 4.857%, its highest level since November 2023, the single most important number here because it is the discount rate attached to every risk asset on the board. Meanwhile the Fed's preferred measure, the PCE price index, is running at 3.7% over twelve months and 4.1% over six months, which is the re-acceleration signal nobody wanted to see. Payrolls added 162,000 jobs in August and unemployment held at 4.1%, so the labour market gives the Fed no reason to look through this.

Which is why Friday's CPI is the only number that matters from here. The Dow Jones consensus is looking for a 0.4% monthly increase and 3.4% over twelve months, with core expected around 0.2% to 0.3%. Natixis' US chief economist Christopher Hodge has framed the whole trade: something below 0.2% on core is probably needed to avoid a hike in September. Anything at or above consensus and the hike conversation becomes a hike decision, and every asset class gets marked again.

Gold has already answered the first half of the question. Spot bullion slid to around 4,340 dollars an ounce within hours of the PPI release, after trading at 4,372 on Wednesday and 4,429.89 on Tuesday. The mechanics are simple: gold pays no coupon, so when the market prices a hike, the opportunity cost of holding it rises and the metal gets sold. What holds it up is the other side of the trade, dollar debasement, fiscal deficit anxiety and a genuine shooting war in the Middle East, with Houthi strikes on Saudi cities, US strikes on Iranian tankers and an Iranian strike on a US base in Jordan all landing in the same week. So gold is locked in a tug of war. The range to watch is roughly 4,250 to 4,300 on the downside and 4,500 to 4,600 on the upside, which is also where strategists expect it to settle.

US equities are the slowest of the three to react because they only trade in cash hours, but they are also the most crowded. On Wednesday the Dow lost just over 400 points, down 0.8%, the S&P 500 fell 0.5% and the Nasdaq dropped 0.6%, a third consecutive down day. The S&P 500 now sits roughly 2% below its August 13 record close while still up about 12% year to date, while futures tried to stabilise before the release, Dow E-minis up 108 points or 0.21% and S&P 500 E-minis up 9.75 points or 0.13%. Earnings are the cushion: 86% of the 492 S&P 500 companies that have reported beat estimates against a long run average of 67.5%, which is why HSBC lifted its year end target to 8,100 and Barclays to 7,950. But the forward multiple is around 20.4 times against a ten year average near 18.9 times, and with the risk free rate at 4.857% that cushion compresses fast.

Crypto is where the answer gets interesting, because it is the only one of the three that never closes. Bitcoin is trading near 77,041 dollars, down 2.91% on the day, with a market cap of about 1.565 trillion dollars and a twenty four hour range between 76,694 and 79,464, close to a 3.6% swing. It is down 2.23% on the week and about 5.4% below the 81,428 high from September 4. Ether is at 2,427.93, down 3.21%, with a 300.47 billion dollar market cap and a range of 2,405.94 to 2,514.24. Solana is at 99.53, down 4.44%. XRP is at 1.3636, down 4.74%. BNB is at 709, down about 5.3%. Dogecoin is the worst of the majors at 0.08379, down 7.71%. Cardano is at 0.21035, down 4.2%. Chainlink is at 11.689, down 3.82%. Avalanche is at 7.606, down 4.74%. Hyperliquid is at 81.24, down 5.95%. Tron is the outlier at 0.3384, essentially flat at 0.05%, which tells you money is rotating to defensiveness, not leaving the asset class. The whole market is worth 2.743 trillion dollars with 89.55 billion traded in twenty four hours, Bitcoin dominance is 59% and Ether dominance is 11.28%, the Fear and Greed index is 68 in neutral territory and the altcoin season index is 37.

The internals tell you how this move was built. Open interest is 53.57 billion dollars in Bitcoin, 33.66 billion in Ether and 6.30 billion in Solana. Funding is barely positive on Bitcoin at 0.0057% and Ether at 0.0024%, and slightly negative on Solana at minus 0.0012%, so shorts pay longs in SOL and leverage is not euphoric anywhere. Taker flow is seller dominated. Bitcoin's taker buy to sell ratio is 0.91, with 32.34 billion of taker selling against 29.42 billion of taker buying, and Ether is 0.95, with 20.75 billion of selling against 19.70 billion of buying. Perpetual order book depth on Bitcoin averages about 804 million dollars two sided and peaked near 833 million, so there is liquidity to absorb size, but the tape is one sided. Hourly RSI is 26 on Bitcoin, 25 on Ether and 25 on Solana, all deeply oversold, and Bitcoin trades below its hourly moving average cluster around 77,900 to 79,100. Spot Bitcoin ETFs saw a 120.24 million dollar net outflow on September 9 after 46.65 million out on September 8, against a 174.6 million dollar inflow on September 4, with total ETF assets at 99.33 billion and 2.05 billion traded. Ether ETFs took in 34.75 million with total assets at 15.69 billion. Crypto stopped behaving like a hedge and started trading like the highest beta expression of the liquidity trade, and the ETF flows confirm it.

So which moves first. My answer is crypto, then gold, then stocks, and the price action this week already proved it. Bitcoin ground lower for three sessions ahead of the print, from 79,656 on September 4 to 78,309 on September 9, because crypto is the only market where a macro repricing can be expressed at three in the morning. Gold was the fastest to react to the actual headline, sliding to 4,340 within hours, because it is the purest read on real yields. Stocks are last because they need the cash session, and they are also the most insulated because earnings are still beating. Rank by speed of repricing and it is crypto first, gold second, equities third. Rank by signal quality and gold is the honest one, because it prices one variable, while crypto prices that variable plus leverage plus sentiment plus weekend liquidity gaps. That is also why crypto overshoots both ways, so a rally off a soft CPI should be bigger in percentage terms in Bitcoin and Solana than in anything equity investors own.

On the second question, position ahead of CPI or wait for it, my honest answer is a bit of both, and the reason is asymmetry rather than a prediction. A hot CPI, headline at or above 0.4% and core at 0.3% or more, almost certainly locks in the September hike. In that world the ten year goes through 4.9%, gold tests 4,250 to 4,300, equities give back another 2% to 4%, and Bitcoin has a clean shot at the 76,700 low and then the 74,000 to 75,000 zone where the next real demand shelf sits. A soft CPI, core at 0.2% or below, does the opposite, violently, because we are oversold on the hourly, funding is flat to negative, and everyone who wanted to be short already is. In that scenario Bitcoin reclaims 79,500 to 80,000 and takes a run at the 81,400 high, Ether goes back through 2,514 and toward 2,600, and Solana is the highest beta bounce of the majors given it is already down 4.44% on the day and 2.7% on the week. So I would carry a partial position into the number, sized so a full stop out costs no more than a normal bad day, with dry powder for the reaction. Pre positioning into a binary macro event is not conviction, it is variance. The people who get hurt on CPI day are rarely the ones who were wrong on direction. They are the ones who were right on direction with too much size.

My broader read is that this inflation fire is not out and the market has quietly moved from debating cuts to debating hikes, which is a regime change, not a data point. That regime, if it holds, favours gold over twelve months more than equities or crypto, because gold is the only one of the three not priced off future earnings growth discounted at a rising rate. But over the next seventy two hours, the fastest and most violent reactions will come from crypto and the most reliable one from gold. Watch the ten year first, watch the dollar second, and treat Friday's core number as the only line in the sand. Levels I am watching: Bitcoin 76,700 and 79,500, Ether 2,405 and 2,514, Solana 98 and 104, gold 4,250 and 4,500, and the ten year at 4.857% and 4.90%. None of this is investment advice, it is just how I am reading the tape into CPI.