alt Hacker NewsI guess you mean informed people. There's a whole lot of smart people in this thread who aren't talking about differentials.
I don't know shit about tariffs, except what everyone said in the news. It was going to tank the economy. Everyone's retirement was going to poof into smoke. Everything was going to cost 2 or 3x previous prices. None of that came to pass.
I bought gas at 2.27 a couple hours ago. Groceries are cheaper than last year. I'm personally making more money than I was a year ago. My business depends on other people having entertainment/spending money.
I'm just a dummy though, so I can't glean some expert insight into the differential. I don't have a model for you, just the real world.
How do you think the economy is doing? What do the differential impact models, that aren't a waste of time, say about it?
It sounds like you are interested, noticing some inconsistencies, and trying to get to the bottom of it, which is a good place to start!
> I bought gas at 2.27 a couple hours ago. Groceries are cheaper than last year. I'm personally making more money than I was a year ago. My business depends on other people having entertainment/spending money.
Comparing Metric(t=0) to Metric(t=1) is a tempting but incorrect way to assess the quality of an intervention. Lots of people think this way, but it is a flawed heuristic that should be avoided whenever possible. Instead, one should compare: PredictedMetric(action=A, time=1) to PredictedMetric(action=B, time=1). This is obvious when one thinks about it, but people get lazy.
To state in another way: when assessing economic policies, it is smarter to compare the observed outcome against the counterfactual outcome. Forgetting or overlooking this is common, but I won't defend or excuse such sloppy thinking.
--
The problem with stating it that way is that I don't think it really drives the point home. Think about someone in a hurry or someone who doesn't know what "counterfactual means... will they stop and _think_? I wouldn't bet on it. So, I'm a fan of hitting people over the head. Show a table:
You can find an example table (with values) at https://budgetlab.yale.edu/research/state-us-tariffs-novembe... (Table 1). For example, it shows that removing tariffs would differentially increase household income by about $900 and reduce employment by 0.3%. That's about 500,000 jobs in the US, assuming a labor force of around 170 million.Will people agree on the models? To say it bluntly (using the "hit them over the head principle"), asking the question like that is bone-headed. Asking it like that is the wrong question, and it misses the whole f-ing point. How on earth we claim to have an educated society when people pose questions like that? We really need to step it up a notch. We've all heard "good journalism gives equal weight to both sides". Fine. But in practice this doesn't get us very far. First, there are more than two sides. Second, hearing out all sides is only the beginning, not the end. Third, practically speaking, if we actually want to make sense of the world in real situations, we're going to discount and possibly completely toss out a whole lot of extraneous, uninformed crap. (Very few serious economists take Trump's economic plans seriously, and there are good reasons for throwing them in the trash! Once one understands what is happening, even mentioning them is usually a waste of time.)
But I digress.
A better question is: on what bases do reasonable people agree and disagree? Using quantitative and substantive models, how can we move forward on making actual testable predictions? Assessing the error in a prediction must not be a matter of opinion. Unless there is a tie, someone is going to be less wrong than the other person. There is no wiggling out of it due to vague language or "miscommunication". That's one key advantage of models. People that care and seek the truth are more likely to share their models. Done right, this will shift the discussion into model specifics and people will have to show their work. This tends to weed out unserious people pretty quickly. (Unfortunately, in many cases, closed models are valuable and so are not shared openly.)
--
I don't often find mainstream journalism that covers any technical topic very well, and this includes economics. I'm not here to blame anyone -- many journalists operate in contexts where time constraints and audience expectations are unlikely to meet my quality bar. Sometimes I will give business analysts a bit more credence, but not much. The world lacks good systems for (a) disseminating clear, testable predictions that (b) lay out their counterfactuals. Heck, I'm surprised when I find even one of the two.
Here's my unsolicited advice. Don't bother reading what "most people" say about economic issues. Ask various friends and network for high quality sources and explore on your own. On this topic, I suggest starting with "When Are Tariffs Optimal?" by Thomas Lubik [1].
Once you have an understanding of what economic models predict, then you can dig a bit deeper. I put close to zero weight when reading mainstream writing on technical topics. If you don't go to the primary sources, you are delegating your thinking. It doesn't take that much work to read the summary from the source material.
[1]: https://www.richmondfed.org/publications/research/economic_b... Federal Reserve Bank of Richmond - Economic Brief - May 2025, No. 25-21