Latest / Stagflation Conversations with Fexingo: High Inflation, Slow Growth, and 1970s Echoes / How AI Productivity Paradoxes Fuel Stagflation
Transcript
- Lucas: Look at the numbers from last Friday’s jobs report and you get a strange picture. Unemployment ticked down slightly to 4.1 percent, but nonfarm payrolls actually shrank by twenty-two thousand. It is a classic stagflation signal where the labor market looks tight on paper, but the actual hiring engine has stalled. Luna: Right, and yet job openings went up to over seven million. So companies are desperate for workers but refusing to hire them because they are waiting for something else to click into place. Lucas: Exactly. And that something is almost certainly artificial intelligence. We have seen massive capital expenditure from the big tech giants this year, pouring hundreds of billions into data centers and chips. But if you look at real GDP growth for the second quarter, it was just one point five percent. That is barely above the starting line. Luna: So the money is being spent, but the productivity boom isn't showing up in the macro data yet. It feels like we are paying for the future while stuck in the present. Lucas: Precisely. This is the productivity paradox all over again, but with a twist. In the nineties, the internet investment eventually lowered costs and boosted output across the board. Today, AI spending is so concentrated in a few sectors that it isn't trickling down to lower consumer prices. Core PCE is still hovering around 336, well above the two percent target. Luna: It is interesting because I keep reading about how AI will cut corporate costs by thirty percent. If that were happening, we should see margin expansion without price hikes. Instead, earnings calls are full of excuses about high energy bills. Lucas: The energy bills are real, Luna. The power demand from these new compute clusters is pushing electricity rates up, which feeds back into inflation. So we have a situation where technology is supposed to be deflationary, but the infrastructure build-out is actually adding to cost pressures. It is a net-negative for inflation in the short term. Luna: That explains why the yield curve is still flashing warnings. Investors know that if growth stays low but prices stay sticky, the Fed can't cut rates easily. Lucas: Correct. And this brings us to the central dilemma for policymakers. The FOMC is watching the ten-year breakeven inflation rate, which crept up to 2.35 percent yesterday. They are afraid that if they don't keep rates restrictive, that expectation becomes self-fulfilling. But keeping rates high crushes small businesses who aren't using AI yet. Luna: So the large firms get to invest in expensive tech that doesn't help them yet, while everyone else pays the price of borrowing money. It seems incredibly inefficient. Lucas: It is inefficient, but it is also predictable. We are seeing a divergence between the stock market and the real economy. The S&P 500 is near seventy-six hundred, driven by those same tech giants. But the Russell 2000 is down two percent over the last week. The wealth effect is not reaching the main street. Luna: I think the key here is the lag time. You can buy a GPU today, but integrating it into your workflow takes months. Maybe years. We are in the messy middle part of that curve. Lucas: Yes, and during that messy middle, you have high capex and low output. It is a drain on cash flow. Look at the average hourly earnings. They are flat at thirty-seven dollars. Workers aren't getting raises because their employers believe AI can handle the next layer of work without human intervention. Luna: But if workers aren't getting raises, consumption slows down. Does that not eventually force prices down? Demand destruction is a powerful deflator. Lucas: It does, but we haven't seen the mass layoffs yet. Initial jobless claims dropped to two hundred and three thousand. Companies are hoarding labor even though they don't need it right now. They are betting that the AI transition will require a different skill set, so they are keeping people on payroll to retrain them. Luna: That sounds like a very expensive way to bet on the future. If the ROI on these AI models is delayed, their balance sheets could take a hit. Lucas: They are already taking hits on interest expenses. Debt service costs are rising faster than revenue growth for many mid-cap firms. This is why we are seeing such a bifurcation in the market. It is not a broad-based recovery; it is a narrow rally powered by speculation. Luna: Speaking of speculation, do you think listeners who are worried about this economic drag should adjust their portfolios, or is this just noise? Lucas: It is real noise, but it is also a signal. If you are looking at this from a career perspective, the advice is to focus on skills that complement AI, not compete with it. Data literacy, strategic oversight, and creative problem-solving are becoming more valuable than routine execution. Luna: That makes sense. You want to be the pilot, not the passenger. But what about the average person trying to manage their household budget in this environment? Lucas: Stick to fixed-income assets where possible. With inflation sticking around three percent, cash loses value, but bonds offer some protection if you lock in rates now. Just avoid long-duration tech stocks unless you have a high risk tolerance. Luna: Solid advice. Before we go deeper into the industrial impact, I want to pause for a moment. Walking through the economy with us has made something click for a lot of our listeners. Lucas: We deliberately don't run ads on these shows. If you want to support that choice, the link is buy me a coffee dot com slash fexingo. Luna: It keeps the research independent and allows us to dig into data like this without pressure from sponsors. We appreciate it. Lucas: Back to the industrial side. The manufacturing PMI has been soft for months. Why aren't factories benefiting from cheaper automation? Luna: Because the upfront cost is too high. A small manufacturer can't just plug in an AI model. They need hardware, software, and training. The barrier to entry is huge. Lucas: Exactly. So the benefits are accumulating only at the top of the food chain. This reinforces the inequality we discussed earlier. The gap between the ai haves and the ai have nots is widening. Luna: It reminds me of the early days of retail banking. The branches closed, but the efficiency gains didn't help the local community until much later. Lucas: A good analogy. And just like then, the transition period is painful. We are seeing supply chains struggle to adapt to new digital workflows. Errors are increasing, not decreasing, in some sectors. Luna: So the promised efficiency is actually creating friction right now. That is counterintuitive but logical given the learning curve. Lucas: It is. And this friction contributes to the sticky services inflation. When your software crashes or your logistics algorithm fails, someone has to fix it manually. That is labor-intensive and expensive. Luna: So the 'services' part of CPI remains stubbornly high because humans are still doing the heavy lifting behind the scenes. Lucas: Precisely. Until those roles are fully automated, which might be five years away, we are stuck with this hybrid inefficiency. Luna: Five years is a long time in politics. Do you think the political cycle will interfere with the economic adjustment? Lucas: Absolutely. Politicians hate stagflation. It is their worst nightmare. We might see populist policies aimed at protecting jobs, which could slow down the AI adoption further. Luna: Regulation as a brake on innovation. That adds another layer of uncertainty to the forecast. Lucas: It does. But remember, the underlying trend is irreversible. The technology is too powerful. The question is just how much pain we endure on the way there. Luna: And whether the GDP figures finally catch up to the hype. I am curious if you think we will see a sudden jump in productivity once this integration phase ends. Lucas: I suspect a sharp V-shaped recovery in productivity metrics, followed by a period of normalcy. But until then, expect volatility and slow growth. Luna: Slow growth and high costs. It is a tough mix for anyone relying on wage increases to keep up. Lucas: It is. So the best strategy is patience and adaptability. Keep your skills sharp and your debt low. Luna: Good advice. As we wrap up, what is one thing you think most people are missing about this ai stagflation dynamic? Lucas: They are focusing too much on the stock market performance of tech stocks and not enough on the real-world implementation costs. The chart on Bloomberg is not the economy. Luna: Right. The reality is in the factory floor, not the trading floor. Lucas: Exactly. And that is where the true story of this decade is being written.