Hacker News Reader

Comment by ajross | Hacker News Reader

Uh... and why would such a correlation be notable? I mean, at the peak of expansions[1] everyone likes to predict recessions. That's literally the same logic as noting that people in the tropics tend to search "hurricane" in August and concluding that the resulting storms are Google's fault.

[1] Check the stock market or some other favorite grown metric. Things hit a peak level about a year ago and have been mostly flat ever since. This is what the peak of a growth cycle looks like; every one can see that.

View on HN

What? Your hurricane example is not the same at all. No one is saying that predictions of a recession in the media are the cause of the recession.

The OP is just asking if there is a way to predict the likelihood of a recession by using media reports. The answer could very well be no, but it is nothing like saying that google searches for hurricanes cause storms.

The correct analogy is that it would be like searching for media reports of incoming hurricanes in order to tell if there was a hurricane coming. I imagine that method would work well for telling if a hurricane is coming (though it would be more straight forward to just look at the weather report)!

I have a feeling this method wouldn't work well for predicting a recession though.

> No one is saying that predictions of a recession in the media are the cause of the recession.

That was the clear implication I took from the phrasing. And it seems to be confirmed by a bunch of other posters here.

Again I ask: everyone can see that we're at the peak of a growth cycle, why is it notable or surprising that pundits are predicting that we'll hit the downward slow soon? It seems it's only notable if you're trying to blame the recession on the pundits.

It would only be notable if pundits can accurately predict if there will be a recession in the next year. I doubt they can because these things are difficult/impossible to predict (otherwise no one would have lost money in the Great Recession in 2008).

You say it is obvious that we are at a peak of a growth cycle. Maybe you are right, but you also could very well be wrong... People have been saying we are at a peak for several years now.

My point was mainly that your analogy was a stretch.

None of that is correct, though. Plenty of people in 2007 were complaining about overheated real estate markets (though no one outside the banking industry knew the disastrous leverage that was in flight). Plenty of people in 1999 (even the Fed chariman! Remember "irrational exuberance"?) could see that the dot com boom was unsustainable. In both situations wehre were predictions of downturns everywhere in the media, and they were right.

Likewise lots of people were predicting a few years solid growth starting in 2009. And they were right too!

And all I'm saying is that it's more likely that they were right because economic cycles are broadly predictable on the multi-year scales and not because Paul Krugman has some kind of magical power to produce a recession just by talking about it.

It would show an actual predictive power of such prognoses.

Contrary to your assertion, I don’t necessarily see a clear signal in the stock market (which only loosely correlates with the economy) or other economic indicators. Job growth is still somewhat strong, for example. And many a temporary flattening of indicators have given way to another run.

Of course as day follows night, recession has followed expansion. But the cycle just isn’t as regular as the 24h day, or the 12 months that make a year.

Just as we have become better at predicting the weather or the path of hurricanes, economists have become better, but not perfect, at predicting, and even softening the blow of, recessions.

OP was merely asking if news archives could be used for a sort of crowdsourced signal, a valid and interesting question. Google trends might also be interesting to look at, although I haven’t heard much of their flu prognosis in recent years.

Bad analogy. There is a plausible mechanism by which consumer sentiment—and, thus, the economy as a whole—can be influenced by media reports. There is no similar mechanism by which the weather can be influenced by media reports.

Bad logic. The presence of a plausible hypothesis doesn't prove causality! This mistake is like 80% of bad science. How many times are we going to repeat it?

The counterfactual I provided is just an obvious example of the fallacy. The real meat of the argument was in that post too -- it's far more likely (Occam and all) that people are predicting recessions because every dummy can see that a recession is likely following the end of an expansion than that there is some kind of magical effect on public thinking.

This is just politics. You don't want your favored entity to get blamed for a recession, so you find a way to make that recession the "fault" of your political enemies. But no, growth cycles are cycles for a reason, and we don't need "fault" to describe what happens.

I was commenting strictly on your analogy. I agree that correlation does not equal causation.

The entire point of the analogy is that it is well known that sentiment does not affect the weather.

It's not a bad analogy, it's an alternate explanation of how you could see the same effect without any causal relationship between recession sentiment and actual recession.

Another example are quack cures: unusual cures are tried when people are at their sickest, then people improve, and they credit the improvement to the cure. Almost everybody who is very sick gets better, though, and the rest are too dead to denounce the cure as fake.

Reply on news.ycombinator.com