On Wednesday, July 29, 2026 the Kospi, the index of the Seoul stock exchange, lost 5.98% in a single session, closing at 5,663.24 points. It was not just any bad day: it was the second day of a historic collapse, after the −10.84% of July 28, with circuit breakers triggered for two consecutive sessions — never seen before — and the index by then around 40% below the record set in June. And yet, even after all this, the Kospi remained around 74% above its level of twelve months earlier. The figures do not contradict each other: over a single day one force is in command, over a year another one is.
The forces that count are four. This guide explains them one by one with the data of late July 2026.
First of all: an index is not “the stock market”
When you read that “the stock market lost 1%”, what is being quoted is an index: a basket of stocks selected according to precise rules, not the entire market. Milan's FTSE MIB contains 40 companies, the Euro Stoxx 50 contains 50, the S&P 500 around 500. Changing the basket means changing the answer.
Almost all the main indices are weighted by capitalization: the weight of each company depends on the value of its tradable shares, and in a concentrated market a few companies decide for everyone. On July 29, 2026 SK hynix gave up 9.61% and Samsung Electronics 5.23%: given their weight, two stocks alone accounted for much of the almost six points lost by the Kospi. Not all of them, though: that day the selling was across the board, and the KOSDAQ — a separate market, where neither Samsung nor SK hynix is listed — closed at −6.12%, with the Korea Exchange suspending trading on both markets. Anyone holding an index fund on that market, theoretically “diversified”, did not dodge it.
| Index | Stocks | Level (July 29, 2026) | 12-month change |
|---|---|---|---|
| S&P 500 (United States) | around 500 | around 7,400 points | +16.2% |
| Euro Stoxx 50 (euro area) | 50 | 6,262.70 points | +16.1% |
| FTSE MIB (Milan) | 40 | 51,443 points | +23.6% |
| Kospi (Seoul) | the entire market | 5,663.24 points | +74.0% |
The distance between Milan and the Euro Stoxx 50, with the same currency and the same monetary cycle, depends on the sector composition of the baskets. And the indices quoted by the media are almost always price indices, which do not count dividends: from January 1957 to June 2026 the S&P 500 returned 7.68% a year on price alone and 10.69% a year with dividends reinvested (6.81% net of inflation). Three points of difference which, over seventy years, make up most of the result.
Force 1: corporate earnings
Over the long run the price of a share follows profits: owning a share gives you the right to a slice of future earnings.
As of July 24, 2026, according to FactSet Earnings Insight, S&P 500 companies were reporting earnings growth of 37.9% year on year: the highest rate since the third quarter of 2021 (40.3%). Revenues were growing by 13.2%, the fastest pace since the second quarter of 2022 (13.9%), and six sectors out of eleven had double-digit earnings growth, with health care the only one falling.
That 37.9% needs to be read properly. A large part of it comes down to a single company: Alphabet reported earnings per share of 9.11 dollars against the 2.88 expected, but inside there was a capital gain of 98 billion dollars. Excluding Alphabet, the index's growth falls to 25.9%: twelve points for a single extraordinary item. Earnings inflated by disposals or revaluations say nothing about the ability to make money selling what you sell: the number to look at is operating profit.
Force 2: interest rates
They are the force most underestimated by savers. At the end of July 2026:
| Reference | Level | Date |
|---|---|---|
| Fed, target range on federal funds | 3.50%-3.75% | confirmed on July 29, 2026 (9-3 vote, three dissents in favor of a hike) |
| ECB, deposit facility rate | 2.25% | July 23, 2026 |
| 10-year US Treasury | 4.63% | July 29, 2026 |
| 10-year German Bund | 3.14% | July 29, 2026 |
| 10-year Italian BTP | 3.98% | July 29, 2026 |
The first channel is discounting. A share is worth the sum of future earnings brought back to today, and to bring them back to today you divide by a rate: the higher the rate, the less tomorrow's profits are worth. One hundred euros of earnings expected in ten years' time are worth 74.41 euros discounted at 3% and only 63.59 euros discounted at 4.63%, the yield on the US ten-year as of July 29, 2026. Almost 15% less without the company having sold a single product less: this is why growth stocks suffer more than the others.
The second channel is competition. When bonds start yielding again, they draw capital away from equities. The comparison is made between the yield on a government bond and the inverse of the index's price/earnings ratio. As of July 24, 2026 the S&P 500 was trading at a forward P/E of 20.1: the inverse is 4.98%, against the 4.63% of the American ten-year. A gap of 0.35 points: very little margin for error.
Then there is an element that is entirely European: the spread. As of July 29, 2026 the differential between ten-year BTPs and Bunds was around 84 basis points (3.98% against 3.14%), and for Piazza Affari it weighs on the cost of credit and on banks' balance sheets.
Force 3: expectations, not facts
It is the concept that most disorients those approaching the markets: the stock market does not react to the news, but to the gap between the news and what was already in the price. As of July 17, 2026, 88% of the S&P 500 companies that had already published their accounts had beaten earnings estimates (five-year average: 78%) and 85% had beaten revenue estimates (average: 70%). As of July 24, positive earnings surprises were at 86%. Beating expectations is the norm, and the usual explanation is that the bar is lowered as the quarter goes on: over the past five years earnings estimates have been cut by an average of 2.0% during the quarter (2.7% over the past ten). In the second quarter of 2026, however, the opposite happened: between March 31 and June 30 analysts raised the earnings per share estimate by 3.4%. The positive surprises came despite a higher bar, not a lower one. This is why “better than expected” is not enough: it has to be much better.
The textbook case is that of July 29, 2026: SK hynix published the best quarter in its history — a fifth consecutive record for operating profit, a 76% margin — and the stock lost 9.61%, because the consensus was expecting around 64,000 billion won of operating profit and 60,543 came in: 5.4% less. The right question is not “is the news good or bad?”, but “is it better or worse than what was already priced in?”.
Force 4: the macro picture and liquidity
Inflation, growth, energy and geopolitics are the tide that raises or lowers almost all the boats together. In June 2026 annual inflation was 2.8% in the euro area (down from 3.2% in May, with a low of 2.4%) and 3.5% in the United States (down from 4.2%, with a low of 2.6%). In the minutes of June 16-17, 2026 the FOMC members flagged upside risks that were still elevated, citing commodities and demand for artificial intelligence infrastructure; some saw reasons to raise rates. On July 29, while confirming the range at 3.50%-3.75%, three members — Beth Hammack, Neel Kashkari and Lorie Logan — voted against, preferring a 25 basis point hike. On July 23 the ECB, leaving them unchanged, listed energy prices linked to the Middle East among the factors to watch.
The stock market is not the economy
In the first quarter of 2026 euro area GDP grew by 0.3% year on year, down from 1.2% in the fourth quarter of 2025. And yet in the twelve months to July 29, 2026 the Euro Stoxx 50 gained 16.1%: indices gather multinationals that make most of their profits outside the borders within which GDP is measured, and they discount the next two or three years.
How much the market costs: reading the multiples
| Indicator | Value | Historical benchmark |
|---|---|---|
| 12-month forward P/E, S&P 500 (July 24, 2026) | 20.1 | 5-year average 19.9 — 10-year average 19.0 |
| Shiller CAPE, S&P 500 (July 28, 2026) | 40.57 | historical average 17.40 — median 16.11 |
The forward P/E divides the price by the earnings expected over the following twelve months: it said that the American market was expensive, but not off the scale compared with the past five years. Robert Shiller's CAPE divides it by the ten-year average of inflation-adjusted earnings, and it was showing more than double the historical average. They diverge because the first incorporates expected growth, exceptional in 2026, while the second compares it with ten years of history. Neither of the two is any use for market timing: the CAPE has been above its own historical average without interruption since the summer of 2009 — the last month below the 17.40 mark was July 2009, and the low (13.32) dates back to March 2009. They are thermometers, not fire alarms.
The two forces people forget: volatility and the exchange rate
The VIX measures the volatility expected on the S&P 500 over the following thirty days and rises when uncertainty grows, which is why it is called the “fear index”. On July 29, 2026 it was quoted at 18.89, within a 52-week range between 13.38 and 35.30. It says how much the market is paying to insure itself, not where prices are going.
Then there is the exchange rate: anyone buying an S&P 500 fund from Milan is also buying dollars. On July 29, 2026 the euro was worth 1.1378 dollars, almost the same as twelve months earlier (−0.44%). But if the index gained 10% in dollars while the euro strengthened from 1.1378 to 1.2378 dollars (+8.8%), the return in euros would fall to around 1.1% — and the mechanism also works the other way round. This is why “currency-hedged” share classes exist, and they have a cost.
Noise and signal: what thirty years of data say
In the short run news, positioning and liquidity are in command: noise. In the long run earnings are in command: signal. According to an analysis by Hartford Funds on Ned Davis Research and Morningstar data updated to March 2026, in 1996-2025 an investor who had stayed invested in the S&P 500 while missing the 10 best days would have had a halved return; missing the 30 best ones, 84% lower. Thirty years wiped out in thirty sessions.
The decisive figure, however, is another one: 76% of the best sessions fell during a bear market or in the first two months of a bull market, that is, when selling seems the most sensible thing to do. It is past history that guarantees nothing about the future, but it explains why getting out “until things calm down” proves costly with remarkable regularity.
Reading a day on the stock market
- Which index am I looking at? Forty Italian stocks and five hundred American ones do not answer the same question.
- Is the move broad or narrow? An index that loses 2% with two stocks plunging and the others flat is not a market going down: it is a piece of company news.
- What was already in the price, and have expectations on rates changed? That is often where the real cause of an inexplicable session lies.
- Does anything change over the next five years? If not, that session is noise.
What to watch in the coming months
- The central banks. With federal funds at 3.50%-3.75%, confirmed on July 29 with three formal dissents in favor of a hike, and the ECB at 2.25% on deposits but alert to the energy shock, in 2026 the risk was no longer only that of postponed cuts: it was that of fresh tightening.
- Spending on artificial intelligence. A large part of the earnings growth of 2026 depended on a few large buyers of infrastructure, and the Fed minutes linked that demand to price pressure.
- Energy and geopolitics. It is the channel through which an external shock feeds back into inflation, then into rates, then into the present value of future earnings.
Stock markets move on earnings, rates, expectations and the macro picture, with the exchange rate as a fifth ingredient for those investing outside their own currency. Telling the noise of a single session from the signal of a decade remains the most valuable skill.
Sources: ECB, “Monetary policy decisions” (23/07/2026); Federal Reserve, FOMC statements of 17/06/2026 and 29/07/2026 and minutes of the meeting of June 16-17, 2026; FactSet, “S&P 500 Earnings Season Preview: Q2 2026” for the upward revision of the estimates, and Earnings Insight (17 and 24/07/2026) for earnings, revenues, positive surprises, forward P/E and the Alphabet figure; Trading Economics (29/07/2026) for the levels of the indices and the twelve-month changes, the ten-year yields, the euro-dollar exchange rate, June 2026 inflation and euro area GDP; Borsa Italiana (29/07/2026) for the FTSE MIB; Investing.com (29/07/2026) for the VIX; multpl.com on Shiller data for the CAPE as of 28/07/2026 and for the monthly historical series; Official Data Foundation on Shiller series for the long-term returns; Hartford Funds on Ned Davis Research and Morningstar data (March 2026) for the best days of 1996-2025; TradingKey (29/07/2026) for the SK hynix and Samsung session and for the fall from the June record; Seoul Economic Daily (29/07/2026) for the KOSDAQ close. The BTP-Bund spread is calculated as the difference between the yields quoted; the levels of the European markets are intraday readings.
Disclaimer: the information in this article is for purely informational and educational purposes and does not constitute personalized financial advice. Past returns are not indicative of future ones. We recommend consulting a qualified professional before any investment decision.