Last night's data shook the market upside down again. The non-farm payrolls for August were only 22,000, while the expectation was 75,000, a significant gap. The June data was revised to a negative number, changing from +27,000 to -13,000, and the unemployment rate also rose from 4.2% to 4.3%.

I haven't seen such data for a few years. I was a bit confused last night; it seems like a 50 basis point cut is on the way, indicating that the economy is quite weak and the expectations for a pullback are strong. The recession I mentioned earlier is not priced in, and the probability of interest rate cuts three times this year is also increasing.
Ideally, bad news drives interest rate cuts, which should be a significant boon. The bond market indeed surged, gold also soared to new highs, and the dollar plummeted sharply. However, the reaction of the US stock market is quite strange, with AMD and Nvidia experiencing significant drops. This is the topic I want to discuss: why is it that the market interprets the same data so differently?
In the first few years after I entered the market, non-farm payroll was one of the most hardcore data points.
Non-farm payrolls good → Indicates a strong economy, companies hiring, strong consumer spending, stock market rising, dollar rising, bond market falling.
Non-farm payrolls → Indicate a weakening economy, increased risk of recession, stock market declines, the dollar falls, and the bond market rises.
Data reflects the economy, and the economy reflects the market. So if this data had come out ten years ago, the market would definitely panic.
The job market is so bad → the economy is going to decline → the US stock market is a sea of red, the dollar is weakening accordingly, and the bond market is surging due to safe-haven demand.
But the past two years have been completely different. After the pandemic, the market has completely transformed, and the core contradiction is the level of interest rates.
This point is easy to understand. In 2020, when the pandemic broke out, the Federal Reserve took extreme measures to inject liquidity, cutting interest rates to zero and implementing unlimited QE, causing the market to surge wildly with too much money. Then in 2022, inflation spiraled out of control, and the Federal Reserve drastically raised interest rates from 0 to over 5%, instantly draining liquidity, and the stock market plummeted from its peak.
So over the past few years, the market has been focusing on one thing: when will the Federal Reserve pivot?
The logic has become:
Data is good → The Fed has no reason to cut interest rates, high interest rates → The stock market is under pressure, the dollar strengthens.
Data deterioration → The Federal Reserve may need to intervene → Expectations for interest rate cuts rise → Stock and bond markets surge, and the dollar weakens.
As a result, a magical outcome emerged: bad news turned into good news.
Last night's non-farm payroll was a textbook case. The data was terrible, and the market's first reaction was not that the economy was finished, but that the Fed would definitely cut rates in September, gold went up, US Treasury yields fell, and the dollar dropped.
But the problem is, the stock market hasn't kept up. Because the data is absurdly bad, it makes people worry that this isn't a slowdown, but a hard landing. This is what I mentioned before: bad news can be packaged as good news, but when it gets bad enough, the logic will reverse.
Here we need to mention a research report viewpoint from Goldman Sachs: the role of the US dollar is switching.
In the past, the strength of the US dollar largely depended on the US economy itself. A strong economy means a strong dollar; a weak economy means a weak dollar. This is straightforward logic, nothing fancy.
But in the past two years, the dollar has been more like a derivative of the Federal Reserve's policy.
Weak data → Market expectations for Fed easing → Weaker dollar;
Strong data → Market worries about the Fed continuing to tighten → Strengthening of the dollar.
In other words, the US dollar has changed from a barometer of economic strength to a barometer of policy expectations, and last night's market movement is a validation of this.
For me, this change is particularly critical, meaning that a lot of old logic has become obsolete. In the past, looking at non-farm payrolls was about assessing the health of the economy; now, it's about betting on policy expectations.
Ok, back to last night, why did gold and the bond market react strongly while the stock market couldn't move?
My interpretation is: For the bond market, interest rate cuts are a huge benefit, and yields will directly decline significantly.
For gold, the logic is also straightforward: the dollar collapses, and real interest rates decline.
But the stock market is different. When non-farm data is extremely bad, corporate earnings expectations will be cut, especially in industries like chips that are highly sensitive to demand. Valuations drop, and as soon as Broadcom's earnings report is released, Nvidia and AMD are pushed down; this is the economic reality.
So the stock market is stuck in the middle: on one hand, the expectation of interest rate cuts is favorable for valuations, while on the other hand, the expectation of a recession is weighing on earnings. The market doesn't know which side to believe, so it moves hesitantly.
My conclusion now is that the market logic has changed.
Traditional logic: Non-farm data → Economic warmth or coldness → Market pricing.
The logic of the past two years: Non-farm data → Policy expectations → Direction of liquidity → Market pricing.
This is also why the same data can be interpreted completely differently by the market at different stages.
Last night was typical: poor data → high certainty of interest rate cuts → bonds and gold celebrating; but to an extreme degree → recession shadow suppressing the stock market → US stocks reacted mildly.
This indicates that the market is currently swaying between two dimensions: liquidity support vs recession concerns.
What I am more concerned about is if in the future this situation arises: employment collapses and inflation remains high. At that time, the Federal Reserve would want to help but wouldn't be able to, as there would be limited room for interest rate cuts. Bad news would signify a recession and wouldn't lead to easing. By then, bad news would revert back to being purely bad news.
In other words, the bad news can still be packaged as good news by the market now because everyone is betting that the Fed has room to save, but if one day that room runs out, the market will truly panic.
It used to be a straight line, but now it’s all twists and turns. Yesterday's situation illustrates this best: the same data, under different logical frameworks, can lead the market to interpret completely opposite results.