专题研究月报回到工作中…直到Xm。..-117600823

September 2025 Jim ReidHead of Global Economics & Thematic Research(+44)-20-7547-2943jim.reid@db.com Henry AllenMacro Strategist(+44)-20-7541-1149henry-f.allen@db.com Rajsekhar Bhattacharyyarajsekhar-a.bhattacharyya@db.com Charts of the month When the Fed is in a rate cutting cycle when there’s no recession, the S&P hastypically seen very strong performance.... ... However US valuations are extreme and 10yr returns from this startingpoint are usually poor, especially in real terms... Contrast this to whenvaluations are low... Perhaps the next few months and quarters will be determined by how theFed is influenced by outside pressure and the data…. 30-year yields are hitting decade plus or multi decade highs in several countries. Isthis inflation, central bank independence or fiscal fears? Likely a bit of all three... While this chart might be too extreme it shows the potential inflation risks anddirection of travel... However the Fed does have a dual mandate so this Friday’s payrolls along with nextThursday’s US CPI will be absolutely crucial… ... Remember that we are hitting a seasonally more challenging period.. The seasonals flip in September and October from lower yields (July andAugust) to higher ones…. When we look back on today’s US equity marketswill it be obvious they’re a bubble, or does thisbull market have further to run…? We look at both sides of the argument… 1. Why this looks like a bubble It’s five years since the government bond bubble burst. It looks obvious inhindsight and waspretty obviousat the time that it was totallyunsustainable. Will a US equity bubble look equally obvious when we lookback on it in five years? When the CAPE has been at a peak, 10yr returns are usually poor, especiallyin real terms. When the CAPE is low the returns are spectacular. Is this timereally any different from 150 years of history? Valuations have usuallymattered, even during this period of US dominance. The top 5 in the S&P 500 are approaching 30% weighting, surpassing the “NiftyFifty” from the late 60s/early 70s and much higher than in 2000… this doesn’tautomatically mean it’s a bubble, but we are in uncharted territory, and it meansperformance heavily depends on a handful of companies If we broaden this out to the top decile of market cap for the entire US stock market,2025 is the most concentrated it's ever been, edging past the early 1930s… Nvidia’s market cap is now larger than every country’s entire listed stock exchangeapart from the US, China, Japan and India. The US is now nearly five times largerthan China in second and around 20 times larger than Europe’s larger markets... What does Warren know that the market doesn’t? Berkshire Hathaway now holdmore cash than they did in the leadup to the GFC… Should this be a warning sign…? Other ratios now even exceed their levels at the height of the dot combubble, like price to sales and enterprise value/EBITDA… The big tech companies have made huge investments in capitalexpenditure… This is a huge high beta play onAIbut can everyone win andmonetise what are becoming huge investments? Could ultimately be betterfor consumers of AI than the producers of it. Are datacentersreally collectively more important than theentire office building sector of the US? Could there be a macro sting in the data center tail? Electricityprices are climbing rapidly…. 2. The argument for higher equities… It could be that macro trumps all valuation concerns. When the Fed have cut ratesinto a soft landing (rather than a recession), the S&P 500 has done very well in thatenvironment… With the rate cutting cycle likely to resume (whether needed or not)we could get a following wind for stocks... Global liquidity is already suggesting the S&P 500 should be nearly 10%higher ResearchDeutsche Bank Remember 1998. Rate cuts after the Russia/LTCM crisis fuelled the last stage of theequity bubble, even when US growth was holding up. Stocks didn't peak for nearly18 months after the September 1998 cut... Many feel that we don't need the ratecuts priced in from September 2025 onwards... In the face of an expected easing cycle, US growth has held upbetter than most feared..... Although the S&P 500’s price-to-earnings ratio is also pretty high byhistorical standards (see also CAPE ratio earlier), it has been high for most ofthe last 30 years. So high valuations alone haven’t been a catalyst for a sell-off even if they might limit the scale of longer-term future returns.... S&P 500 Price-to-Earnings Ratio (red line is 5 year moving average) It’s clear from company earnings transcripts that the AI momentum isn’t letting up…if it doesn’t let up then money will flow into the perceived beneficiaries... If Nvidia delivers on these analysts' forecasts, then we have truly moved to a new eraof even greater US dominance and AI will have cemented supreme USexceptionalism... Longer term much will depend on productivity… If AI can offer a sustained burst ofhigher productivity l