Goldman Sachs' equity strategy team raised its three-month rating on global equities from neutral to overweight, aligning it with its existing one-year bullish stance.
The team's report cited rising oil prices and bond yields as factors that had stalled the prior rally, thereby easing valuation pressure, while corporate earnings prospects remain solid. "Earnings growth continues to support equities, and the recent valuation pullback has improved the risk-reward into the third-quarter earnings season," said team head Christian Mueller-Glissmann.
To be clear about the nature of this upgrade: it is a tactical asset-allocation rating, not a price-target forecast. Goldman's target of 8,000 for the S&P 500 by the end of 2026 (raised from 7,600 in May) is a separate matter, and this report contains no new index-level call. According to the latest compiled broker targets, 8,000 sits roughly in the median-to-optimistic range: BofA Global Research has 7,400, Barclays has 7,950, JPMorgan, Deutsche Bank, Morgan Stanley and Goldman Sachs all have 8,000, while Citi, UBS Global Research and HSBC have 8,100.
The upgrade is a bet on a high-stakes earnings season. According to data compiled by industry research, analysts expect S&P 500 constituent profits to grow 25% year over year in the third quarter, one of the best growth rates on record; Goldman's team believes actual results will exceed that threshold. The same thing measured by a different statistical yardstick changes the height of the bar. According to FactSet, S&P 500 third-quarter earnings per share are expected to grow 29.5% year over year with revenue growth of 12.3%; if realized, that would be the third consecutive quarter of growth above 25% and the eighth consecutive quarter of double-digit growth. Fourth-quarter growth is projected at 27.6%, and full-year 2026 growth at 32.4% (data as of Oct. 2). The gap between 25% and 29.5% is 4.5 percentage points, and which one Goldman intends to "exceed" directly determines the weight of that statement.
This time, expectations were not revised down ahead of earnings season. FactSet noted that from the end of June to the end of September, S&P 500 third-quarter EPS expectations were actually revised up by about 1.4%, versus an average downward revision of 2.2% over the same period in the past five years. As of Sept. 25, 72 of 116 companies that had issued quarterly guidance (62%) gave positive guidance, far above the five-year average of 40%. Even more noteworthy is the concentration of earnings growth. According to an analysis by ZeroHedge citing earnings forecasts from Goldman Sachs and FactSet, Nvidia (NVDA.US) and Micron (MU.US) are expected to account for more than one-third of S&P 500 third-quarter EPS growth, contributing more than the bottom 490 companies in the index combined; AI infrastructure-related companies are expected to contribute more than half of the quarter's EPS growth. Barclays' calculations point in the same direction: semiconductors and tech hardware together account for only 29% of S&P 500 weight, yet they contributed about 75% of fiscal 2027 EPS upward revisions and are expected to contribute about 70% of that year's index earnings growth. Barclays also ran a stress test: if fiscal 2027 semiconductor EPS were revised down 15% while other sectors' earnings stayed flat, the S&P 500 could still achieve double-digit growth above its long-term trend; only if semiconductors and other sectors were revised down simultaneously would index growth likely fall back near its long-term trend. The premise for that buffer is that other sectors do not decline alongside.
Earnings quality also leaves a gap. According to an Oct. 7 note from Charles Schwab, third-quarter tech-sector earnings are expected to grow about 65%, part of which comes from investment gains on holdings such as SpaceX and OpenAI. That portion is paper appreciation, not cash, and could reverse if the AI boom slows.
Valuations already digested: evidence and counter-evidence
Goldman's starting point is that the pullback has lowered valuations. This is supported by data: the S&P 500's forward 12-month price-to-earnings ratio has fallen from 20.4 times at the end of June to 19.0 times, below the five-year average of 19.8 times and roughly in line with the 10-year average of 19.1 times (FactSet, as of Oct. 2). But putting stocks and bonds side by side flips the conclusion. Based on Oct. 9 data, the S&P 500's forward earnings yield was 4.33% while the 10-year U.S. Treasury yield was 5.24%, a gap of -91 basis points, meaning the earnings return from holding stocks is lower than the interest from holding Treasuries. That places the metric at the 11th percentile of its entire history since 1990, versus a historical median of +198 basis points. A rougher measure tells the same story: based on Shiller's long-term data, the simple difference between the S&P 500 earnings yield and the 10-year Treasury yield was -1.42% on Oct. 1, at the 10.3rd percentile, versus a historical average of +2.71%; Shiller's "excess CAPE yield" read 0.56 points in October, the lowest since March 2002. This is the flip side of "improved risk-reward": valuations have indeed pulled back from highs, but they have pulled back to a point where the stock-bond yield gap is near zero. At 7,802 points and forward EPS of about $411, each one-point change in the P/E ratio corresponds to roughly 410 index points, meaning 17 times equates to about 6,981 (-10.5%) and 21 times to about 8,623 (+10.5%). At this level, a small earnings miss would be amplified more than usual.
Barbell, regions, and that Asia report ten days earlier
The allocation structure Goldman laid out is a barbell strategy: one end in global tech stocks, the other in select rate-sensitive "bond-like" stocks whose valuations have already been digested since the summer. Regionally, it maintains overweight on Asia and the U.S., and underweight on Europe. The Asia leg already had a fuller version ten days earlier. On Oct. 8, Goldman published a global strategy report titled "Investing in a differentiated Asia" (Global Strategy Report No. 78), giving a 12-month target of 1,120 for the MSCI Asia Pacific ex-Japan index, corresponding to about 26% price return in dollar terms, and projecting regional earnings growth of 75% in 2026, 25% in 2027 and 14% in 2028, versus a compound growth rate of only 6% for the region from 2011 to 2025. It also raised its long-term forecast: annual earnings growth over the next decade from 9% to 10%, with a 10-year annualized dollar total return forecast of 13.6%, at the 89th percentile historically. The report noted that information technology will account for 47% of 2026 regional earnings, with memory chips potentially making up about 70% of that; the risk is stated plainly in concentration, with the top six tech stocks accounting for 36% of the index's market capitalization. Within the region, Goldman is overweight South Korea, Taiwan, Japan and China A-shares, and underweight Australia, Thailand, Indonesia and the Philippines, with a 12-month target of 12,000 for South Korea's KOSPI. Europe is the other end of the chain. A Sept. 18 survey showed strategists' median year-end target for the Stoxx Europe 600 at 670, the most optimistic reading since 2018, while Goldman maintained an underweight on Europe this time. The two are not necessarily contradictory — one is a target price and the other a regional allocation weight — but the directional difference is clear: Goldman macro trader Rikin Shah judges that the U.S. is clearly an AI winner while Europe is not, and a renewed widening of sovereign bond spreads would further tighten European financial conditions. In intraday European trading on Oct. 12, the Stoxx 600 rose 0.56% to 635.09, leaving about 5% of room to the 670 target.
The real shift is in bonds
More worth reading than the equity rating in this report is the bond call. Goldman kept its three-month rating on bonds unchanged at neutral but raised its 12-month rating to overweight, offering a line rarely seen in sell-side research: "We believe bonds should gradually shift from a drag on portfolios to a hedging tool." Only by reading this alongside the equity call does the logic become complete: being overweight both stocks and bonds over 12 months means Goldman is earning from corporate profits while using bonds to hedge the two risks it itself lists — higher rates and higher energy prices. This also explains why it separates "equities turned overweight on a three-month view" from "bonds still neutral on a three-month view": within three months rates are still rising and bonds are still under pressure, so no hedge; over 12 months, duration begins to become protection.
The other side: BofA is stockpiling cash
Not every sell-side firm is moving in the same direction. Michael Hartnett of BofA Securities, in his Oct. 8 "Flow Show" report, offered a different judgment: the U.S. midterm elections on Nov. 3 are the biggest potential catalyst for a plus-or-minus 10% swing in global stocks before year-end. He cited market forecasts showing the probability of a Democratic sweep of both chambers has risen from 49% in September to 64%, and if that scenario materializes, U.S. stocks face more than 10% downside risk; BofA's bull-bear indicator fell from 8.8 to 8.1, maintaining a "sell" signal. Fund flows are on his side: global money-market funds took in $166.4 billion net that week, the largest weekly inflow since April 2020, with global money-market fund assets rising from $5 trillion in 2023 to $8 trillion; U.S. Treasuries saw $16.6 billion in net inflows; financial stocks suffered their largest weekly net outflow since March 2026, about $3 billion. BofA Global Research's year-end 2026 S&P 500 target is 7,400, among the most cautious forecasts on Wall Street, and analyst Savita Subramanian, while raising it from 7,100 to 7,400 in September, still flagged interest-rate risk. Historical midterm election statistics provide ammunition for both sides. According to LPL Financial, in 17 midterm elections since 1954, the S&P 500 rose in all 17 in the 12 months after the vote, with an average gain of 18.2%; but the average maximum drawdown in midterm election years themselves was 19.4%, versus only 12.5% in other years (since 1960). In other words, volatility and returns occur sequentially in such years, not simultaneously absent.
Back to the testable variable: whether this rally can last to year-end depends on whether the 25% to 29.5% earnings bar is cleared or sidestepped. With the S&P 500 just shy of its record closing high and in a valuation structure where the stock-bond yield gap has already hit zero, Goldman is choosing to add equities while BofA is choosing to hoard cash — both are waiting for the same batch of earnings reports.
Comments