One Percent From a Record, at 2007 Prices

The S&P 500 closed Friday at 7,743. That is up 93 points, or 1.2 percent, and it is the highest weekly close in six weeks. The record high is 7,817. The market is now nine tenths of one percent away from an all time high.

Almost the entire gain came on Monday. The index rose 114 points in one session, its best day since before the August peak, on two pieces of news. Chip and artificial intelligence companies surged after the United States and China agreed to an AI dialogue during a three day summit in Washington. And oil fell hard because American and Iranian negotiators were meeting in New York to discuss reopening the Strait of Hormuz.

Then Saturday happened. President Trump rejected Iran's latest proposal. So the market heads into this week priced within a percent of a record, on the strength of a deal that has just been turned down, while the cost of borrowing money sits at the highest level in a generation. That is the setup, and it is not a comfortable one.

Right on Rates. Wrong on What Rates Would Do to Stocks.

Last issue named two scenarios and the market took the shape of one and the mechanism of the other. Scenario A said the low at 7,508 would hold and the market would work higher into the 7,687 to 7,699 zone. Both happened. But it also said the rally would stall below 7,717 and the pattern of lower peaks would stay intact. It did not stall. The market closed above 7,717 on two days and reached 7,782, sixty five points past the ceiling we named. The pattern of lower peaks was not preserved. It was erased. Tuesday's high took out four of the five declining peaks in a single week.

The bigger miss was Scenario B, and it is worth stating plainly. We said a hawkish message would push the 10 year Treasury yield through 5 percent and close the market below 7,610. Yields went far past 5 percent and the market rose anyway. The 10 year closed at 5.184 percent. The 30 year closed at 5.504 percent. The whole curve rose between 15 and 19 basis points. Every part of the rate call was right and the consequence we attached to it was wrong. Rising yields did not sell stocks last week. Own that.

Two calls landed cleanly. We said to watch the first two sessions because the options expiration pinning was gone, and Monday delivered the biggest move of the month. And on the euro we said two matching lows at 1.1459 and 1.1460 were a magnet rather than a floor. They were taken out, and both named targets at 1.1453 and 1.1421 were reached and passed, down to 1.1364. That was the best call of the issue. Gold was the worst. We said it needed only to clear 4,429 to confirm its low. It reached 4,422 on Monday, failed seven points short, and fell to 4,278. The warning level we published, a close back below 4,381, did exactly what we said it would and put the low back in question.

Money Costs More Than It Has in Twenty Two Years

The number that matters most last week was not in the stock market. The 30 year Treasury yield closed at 5.504 percent. That is the highest since June 2004. The 10 year closed at 5.184 percent and touched 5.23 percent during Thursday's session, the highest since 2007. The 5 year closed above 5 percent for the first time since July of that year.

This was not an American event. Japanese, British and German government bonds all sold off at the same time, the second week in a row that has happened. Two things drove it. Federal Reserve officials came out of their quiet period and talked tough. And the data refused to cooperate with anyone hoping for a pause. Weekly claims for unemployment benefits came in at 197,000, the lowest since mid July and among the lowest readings since 1969. The number of people continuing to collect benefits fell by 55,000 in a single week. Nobody is losing their job, which means the Fed can keep raising rates without worrying about employment. Sixteen of its eighteen officials already expect at least one more increase, and only two meetings remain.

Here is where that reaches ordinary life. The 30 year fixed mortgage rate jumped to 7.45 percent on Thursday, the highest in more than two years. Utilities fell 3.0 percent on the week and energy fell 3.1 percent, both squeezed by the move in long term rates. Meanwhile technology rose 3.1 percent. The Nasdaq gained 2.1 percent, the Dow gained 0.3 percent, and small company stocks actually fell 0.8 percent. Read that last one again. Small caps declined in a week the index rose 1.2 percent. A market that only advances in its largest names is narrower than the headline number suggests, and it has been getting narrower for six weeks.

The one genuinely constructive development came from the summit. The United States and China agreed to cut tariffs on 30 billion dollars of goods in a 90 day arrangement, set up a trade council, and open a formal dialogue on artificial intelligence. Neither side has yet said how far tariffs fall or when. Those details are due Monday.

The Record Is Close. So Is the Line That Cancels It.

On the upside the interesting thing is not one number, it is a cluster. Several unrelated measures land between 7,812 and 7,832, and the record high at 7,817 sits inside that band. That zone is the target, and it is only about 1.1 percent away. There is a detail worth noting inside it. This week's projected ceiling is 7,823, which is above the record high. That has not happened before in this letter. It means a new all time high is inside a normal week's trading range from here rather than something the market has to stretch for.

On the downside the number to write on your hand is 7,663. That was Thursday's low, and it is the first higher low this market has made since August. If the recovery that started on the sixteenth is real, the market should not close below it. A close below 7,663 would mean the market failed to hold the low it had just claimed, one week after claiming it. It is 1.0 percent below Friday's close. Below that sits 7,636, the average price of the last fifty trading days, which is rising again at nearly nineteen points a week after almost flattening out the week before. Then 7,595, the monthly floor, which was not tested at all last week. Then 7,498, which has not moved in three weeks and is now 3.2 percent below the market.

Now the caveat, and it is the most important thing in this issue. The level that confirmed the recovery, 7,760, was cleared by four and a half points on Monday and Tuesday and then lost. The week closed 17 points below it. Friday's high came up eight points short of taking it back. A signal worth less than a tenth of one percent, held for two days and then given away, is not much of a signal. The recovery is confirmed on paper and unconfirmed in practice.

Gold is the cleanest decision on the board. It closed Friday at 4,321, fell to 4,294 in early trading, and sits just above two matching lows at 4,278 and 4,273. Those two are five points apart. They have been tested twice and not broken. Two lows that close together are a magnet, and that is the same observation that proved correct on the euro last week. Everything above gold is resistance. The first hurdle is 4,352, then 4,360, and the level where the trend of lower peaks actually breaks is still 4,480. A close below 4,273 opens 4,222 and then 4,133. A third hold followed by a close back above 4,352 would be the opposite trade and a strong one.

The euro is at 1.1392 after falling to 1.1364 on Thursday. It has done most of its work. There is roughly 39 pips of decline left, to 1.1353, before this move has travelled its full measured distance. That is the entire remaining bear case, which is why the euro is not interesting here in either direction. Above the market, supply is stacked between 1.1434 and 1.1488, and a close above 1.1432 would be the first sign the decline is finished.

The Peace Trade Got Rejected on Saturday

The chart is pointing modestly higher, into a record, and it has been pointing that way since Monday. Geopolitics is not what will cause that move. It is what gives the market permission to finish it. Last week the permission came from New York, where American and Iranian negotiators discussed a regionwide ceasefire of up to 60 days, a phased reopening of the Strait of Hormuz, and an end to the American naval blockade, with Qatar, Pakistan and Egypt mediating. On Saturday the President rejected Iran's proposal, and talks are expected to resume this week. The story that paid for Monday's rally has been withdrawn, at least for now.

So the test is simple. If talks resume and make visible progress, the market gets its permission back and the last percent to a record is easy. If they stall, the buyers who paid up on Monday are holding a position built on a headline that no longer exists. One more thing is not yet priced. Saudi Arabia is sending no crude at all to European customers in October and is rerouting roughly 60 million barrels through the Persian Gulf instead. That puts more oil through the exact waterway the damaged pipeline was built to avoid, at a time when the pipeline is still being repaired. A quiet week keeps that invisible. A single strike does not.

The Busiest Calendar Since August, Landing on Quarter End

Last week was the quietest calendar since the middle of August. This week is the opposite. Monday brings the full detail of the China tariff agreement, and since neither government has said how far tariffs fall or when, that is a real event rather than a formality. Tuesday brings consumer confidence and job openings. Wednesday brings the inflation measure the Federal Reserve actually targets, and it is also the last day of the quarter. Thursday brings weekly claims and factory activity. Friday brings the September jobs report, with economists expecting 100,000 new jobs against 162,000 last month.

The risk in that schedule is not balanced, and it is worth understanding why. Borrowing costs already moved 15 to 19 basis points last week. A hot inflation number on Wednesday or a strong jobs number on Friday lands on a bond market that has already travelled a long way and would push the 10 year toward its 2007 high near 5.25 percent. A weak jobs number is the relief, because it is the only thing on the calendar that argues the Fed should stop. There is also a mechanical wrinkle on Wednesday. Large funds rebalance at quarter end to return to their target mix of stocks and bonds. This quarter stocks rose and bonds fell hard, so the mechanical trade is to sell some stocks and buy some bonds. That is a one or two day effect rather than a signal, but it will muddy the reaction to the inflation print.

Two Ways This Week Resolves

Scenario A, which the chart slightly favours. The low at 7,663 holds and the market grinds into the 7,812 to 7,832 band, tagging or clearing the record high at 7,817. The early tell is a close above 7,745 in the first day or two. But understand what that zone is. It is also where the intermediate term setup expects the turn, and it arrives with the market narrower than it was a month ago and small caps already falling. A record high in the first half of October is the setup for the next decline, not the end of the last one. Treat 7,812 to 7,832 as a destination to sell into rather than a base to build on.

Scenario B. The market closes below 7,663, which says the recovery failed the week after it was claimed. That opens 7,636, then 7,595 on a first real test, then Wednesday's low at 7,508. And 7,498 remains the line that would say the advance since March is finished. Below it, the long term average price comes into view, near 7,205 today and rising toward 7,271 by the middle of October.

The honest counterweight to both. The level that was supposed to confirm this recovery was cleared by four and a half points and surrendered within three sessions, and the week closed below it. Meanwhile long term borrowing costs are at a twenty two year high, mortgages are at 7.45 percent, and the peace negotiation that paid for Monday was rejected on Saturday. A market a percent from a record on those inputs is not wrong, but it is leaning on very little.

What to Do With This

The box this week is 7,663 to 7,832. Below the first, the recovery failed. Into the second, the market meets a record high and a cluster of resistance at the same time. Unlike last week, the market is not sitting in the middle of its range. It is near the top of it, which means the risk and the reward are no longer symmetrical.

7,745 is the early tell. A close above it in the first sessions says the market intends to finish the job and go for the record. Failure to clear it while 7,663 holds is chop, and chop near the top of a range is distribution more often than it is accumulation.

Monday is a real event. The tariff detail has not been published and the market has already rallied on the announcement. That is the order in which disappointments happen.

Wednesday and Friday decide the bond market. Listen for the inflation print and then the jobs number. The stock market has spent a week ignoring a generational move in borrowing costs. That has worked for five sessions. It is not a strategy.

Gold has the clearest decision and the euro has the least to offer. Gold is sitting on two matching lows five points apart with every level above it acting as resistance, so it resolves one way or the other and soon. The euro has about 39 pips left in its decline, which is not enough to be worth the risk in either direction.

A final thought. The market is one percent from an all time high while money costs more than it has since 2004. Both of those things are true right now and both cannot stay true indefinitely. The calendar points to the second half of October as the window where that gets settled. Between now and then the market has an invitation to print a record first, and the trap in this market is treating that record as the beginning of something rather than the end of it.