Why an Interest Rate Hike Is Still on the Table for June

WSJ发布于2023-05-22更新于2023-05-24

文章摘要

The fire hose of commentary from the Federal Reserve this past week made one thing clear: Central-bank officials are fiercely split on how to navigate the policy path forward and whether to raise interest rates or hold them steady when the Fed’s policy-making committee next meets in June.

The fire hose of commentary from the Federal Reserve this past week made one thing clear: Central-bank officials are fiercely split on how to navigate the policy path forward and whether to raise interest rates or hold them steady when the Fed’s policy-making committee next meets in June.

Officials laid the groundwork at their May meeting for the Fed to pause next month, although they stopped short of committing to specific next steps. But since then, a series of economic surveys and data releases have appeared to show an economy that looks increasingly stable.

That has emboldened the central bank’s hawks to publicly endorse the idea of at least one more rate boost in June, given that the pace of inflation remains more than double the Fed’s target. Dallas Fed President Lorie Logan, a voting member of the Federal Open Market Committee, was perhaps the most explicit, saying Thursday that while the data “could yet show” it is appropriate to skip a rate hike, the current environment suggests that “we aren’t there yet.”

The economic case for further policy tightening centers on fresh optimism that the U.S. will be able to stave off a recession, at least until sometime next year. A new working paper from economists at the San Francisco Fed shows that, despite years of strong demand and high inflation, U.S. households across the income spectrum still hold an estimated half-trillion dollars in excess savings—enough to fuel consumer spending at least into the fourth quarter of 2023.

At the same time, some of the more interest-rate-sensitive sectors of the economy that had been impacted by monetary-policy tightening—namely, housing and manufacturing—could be reaccelerating.

Skanda Amarnath, executive director at Employ America and a former research analyst with the New York Fed, notes that growth in multifamily residential units under construction and resilience in mortgage demand both are helping to paint an “encouraging picture” for the housing market. Manufacturing data out this past week from the Philadelphia Fed, he added, suggest that the most contractionary months for that sector could be behind us.

Buoyancy in housing and manufacturing isn’t “a typical thing you see right before a recession,” Amarnath said.

There is also the question of whether the regional-banking turmoil of the past few months caused financial conditions to tighten beyond what the Federal Reserve had intended.

Fed Chairman Jerome Powell noted on Friday that rates might not need to rise as high as the central bank had anticipated because of bank stress. But so far, the impact has been less dramatic than anticipated.

The central bank’s own survey of senior loan officers showed bank lending practices tightening only mildly between the fourth quarter of 2022 and the first quarter of 2023. “Honestly, I expected that the turbulence...was going to cause a lot more panic,” Atlanta Fed President Raphael Bostic said at an economic conference in Florida this past week.

Chicago Fed President Austan Goolsbee, also on stage for the keynote discussion, agreed.

Fed officials still have nearly a month before they must make a decision, and Powell on Friday signaled the need for the central bank to tread carefully.

Even so, the recent signs of economic strength suggest that monetary policy could have more room to run. If any outstanding data come in hotter than expected, brace for the possibility of at least one more increase in the federal-funds rate.

你可能也喜欢

三年之后:回看 2023 年我对 ChatGPT 的判断

作者王健硕在2026年5月回看其于2023年3月对ChatGPT做出的二十条预测,并借助AI代理进行验证。整体看,其大方向判断多正确,但细节和程度常有偏差。 **看对的方面:** - **技术架构**:准确预测RAG(检索增强生成)将成为解决知识更新和幻觉的主流方法,以及LUI(自然语言用户界面)将催生巨大新产业。 - **发展趋势**:预见到“机器人网络”和新的agent寻址系统将出现,中国能快速做出可用大模型并缩小与顶尖差距。 - **本质认知**:正确指出ChatGPT没有意识,图灵测试仅测表象;判断其是巨大进步但非AGI,且短期内不会造成整体失业潮。 **看错或看偏的方面:** - **具体数据错误**:关于GPT-4有100T参数的传闻完全错误。 - **绝对化判断**:断言LLM“不可能”自己学好数学被后来模型在IMO夺金证伪;认为AI生成内容可“规避”版权,实则引发史上最大侵权赔偿。 - **价值与成本误判**:认为价值终落应用层,但最大赢家是算力层(如英伟达);模型成本“5-10亿封顶”的估算严重偏离实际。 - **社会影响误读**:认为ChatGPT的“加权平均”特性可能促进“世界大同”,但实际AI正走向个性化,并可能制造新信息茧房。 **总结规律:** 1. 预测机制和方向比具体数字更可靠。 2. 倾向于高估短期变化速度,低估长期影响程度。 3. 容易忽略问题内部的“分布”差异(如失业影响集中在年轻群体)。 4. 留有餘地、分层表述的判断更经得起时间检验。 5. 一些根本性争议,三年时间仍不足以给出定论。 这次复盘旨在为未来的判断立下更审慎的规矩。

marsbit2小时前

三年之后:回看 2023 年我对 ChatGPT 的判断

marsbit2小时前

三年之后:回看 2023 年我对 ChatGPT 的判断

2023年3月,在GPT-4发布前,作者王健硕对ChatGPT的未来做出了二十项预测。三年后的2026年,他使用AI agent对这些判断进行回顾验证,评估其准确性。 **主要看对的方面:** 1. **RAG成为主流**:预测通过外部检索而非修改模型来解决知识更新和幻觉问题,这已成为行业标准架构。 2. **LUI(自然语言界面)兴起**:预测自然语言交互将催生巨大新产业,Agent、MCP协议等发展印证了这一点,但LUI是与GUI共存而非取代。 3. **机器人网络与新寻址系统**:预测Agent将自动协作并需要新寻址方式,MCP、A2A等协议正朝此方向发展。 4. **中国能做出可用大模型**:预测中美模型差距会迅速缩小,事实如此,但真正领先的国内厂商与其当初点名不同。 5. **AI无意识与图灵测试本质**:核心判断“AI无意识,图灵测试仅测表象”基本成立,但“绝无意识”的绝对论断被后续研究置于灰区。 **看错或看偏的方面:** 1. **GPT-4参数数量**:所传“100T参数”严重错误,实际约为1.8T。 2. **LLM的数学能力**:诊断“数学差需外挂工具”正确,但“不可能自行学会”的结论被后续“推理模型”在IMO夺金证伪。 3. **价值捕获**:预测价值在应用层,但现实中算力层(如英伟达)捕获了最大利润,模型层反而亏损。 4. **版权问题**:判断“生成物可能无法登记版权”正确,但认为能“规避侵权”则错误,已出现巨额侵权赔偿案例。 5. **信息茧房与世界大同**:机制上AI会将观点向众数平均,但AI正快速走向“千人千面”的个性化,可能制造新茧房而非消解极化。 6. **大模型成本**:预测“局部战争”和玩家涌入正确,但“5-10亿美元封顶”的成本估算与前沿训练实际花费严重不符。 **总结规律:** 1. 判断**方向和机制**比具体数字、程度更可靠。 2. 倾向于**高估短期变化速度,低估长期能力上限**。 3. 容易忽略**问题内部的分布差异**(如总量正确但部分群体受损)。 4. **留有餘地、分层表述**的判断更经得起时间检验。 5. 一些根本性争论,**三年时间不足以给出最终答案**。 这次回顾表明,在快速变化的领域,把握大方向比追求精确预测更重要,同时需警惕绝对化表述并承认认知的局限性。

链捕手4小时前

三年之后:回看 2023 年我对 ChatGPT 的判断

链捕手4小时前

交易

现货
合约
活动图片