Are Professional Forecasters Overconfident? (newyorkfed.org) 32
Research by the Federal Reserve Bank of New York, published on Tuesday, indicates that professional economic forecasters tend to overestimate their accuracy in long-term predictions while underestimating their short-term precision.
The study, which analyzed data from the Survey of Professional Forecasters from 1982 to 2022, revealed that for forecasts two to four quarters ahead, actual errors were two to four times greater than the forecasters' estimated uncertainty ranges for both GDP growth and inflation. In contrast, for predictions less than three months out, forecasters typically overestimated potential errors.
The study's author, Marco Del Negro, highlighted significant differences in uncertainty estimates among individual forecasters, suggesting that these findings challenge the rational expectations theory. Del Negro proposed that these discrepancies might stem from an over-reliance on varying models or priors in making longer-term forecasts.
The study, which analyzed data from the Survey of Professional Forecasters from 1982 to 2022, revealed that for forecasts two to four quarters ahead, actual errors were two to four times greater than the forecasters' estimated uncertainty ranges for both GDP growth and inflation. In contrast, for predictions less than three months out, forecasters typically overestimated potential errors.
The study's author, Marco Del Negro, highlighted significant differences in uncertainty estimates among individual forecasters, suggesting that these findings challenge the rational expectations theory. Del Negro proposed that these discrepancies might stem from an over-reliance on varying models or priors in making longer-term forecasts.
Do humans easily become overconfident? (Score:4, Interesting)
When people see a computer simulation closely match short term reality they tend to project that this will happen in the long term. Newton produced a good model of physics that models everything quite well until people considered things travelling a significant percentage of the speed of light. People see LLMs answer some questions good and start to assume it will always be good (using Wikipedia for everything without checking sources happened previously). Forecasters see the models work really good over certain conditions and lengths of time and assume it will work for everything.
People will continue to ignore the constraints and accuracy of the data sources they work with, not exactly a new insight.
Re: (Score:2)
Re: (Score:2)
Exactly. Nobody that is good at making prediction wants to do that as a main part of their job. Too depressing.
Remember to 'trust' ellected/appointed officals (Score:2)
How many of these election cycles do we have to endure before people stop pretending that numbers thrown up and projected by congress, presidential candidates, the Congressional Budget Office, World Bank, IMF, WEF, UN are anywhere near accurate.
I think the first question reporters or people should ask these people is "What have you directly done with your own money, your agency budget money to do X in the last 5 years?"
This stream of lazy journalism, lazy economic analysis, lazy economists all seeking to no
Re:Do humans easily become overconfident? (Score:4, Informative)
Niels Bohr: "Prediction is very difficult, especially if it’s about the future!"
Add to this the particular problems with economics.
First, economics has very little models that are actually worth a damn. They cling to unworkable ideas like the utility-maximizing individual as a base model of human behaviour, doing their best to avoid the problem that neither have they any actual definition of utility, nor is there such a maximizing individual anywhere to be found in the wild. Then there's gems like supply-side economics, which it should be obvious to anyone with a functioning brain that it will not work - what good is supply, if there is no one with money to buy? Or trickle-down, come on, literally all of the last 50 years show us it doesn't and cannot work, if there should be anyone that cannot put 2 and 2 together. It's almost as if:
Second, the field is financed by the monied class. Economists flock to the grants and prizes offered, but these are offered to do theory and research that affirms the interests of those offering. In this economists are the only people to actually have an utility they are maximizing - getting as much of that prize money as possible. As a result you cannot really do much of actual economics as an insider in the field, nor much of a career outside of it. The monoculture is such a problem by now that it's inceasingly difficult to put together any kind of initiative to even think about something different. The field has been recreating the same status quo affirming theory, and economists, for generations now, and how's it working for us by now? Anyone happy with the economic situation of their respective nation? As an example, pretty much anywhere in the West, the new generation is never going to be able to afford homes. Do we have any economics offered anywhere to deal with that problem?
Last, even if we assumed that everything was ok with the previous issues, there is also a big garbage-in-garbage-out problem. Start with the simplest of misdirection, e.g. GDP as a measure of economic well-being. Well the GDP is made of the stock market and corporate profits, so ig they are doing well, but are we not forgetting the whole point of it is supposed to be the well-being of people? And would you look at that, real wages have been flat for the aforementioned 50 years. So why do we talk about the first, but not the latter? Then there's the question of cooked books, like jobless statistics - US statistics declares a jobless as someone who receives jobless benefits, as opposed to someone who is, in fact, jobless - turns out there's much more of the latter. No one cares to count the homeless, no one cares to count long covid, and so on. But you cannot make good policy on bad data. This, of course, is only a problem if you actually care if your policy is good or bad wrt actual societal outcomes, a care that is less and less to be found.
Re:Do humans easily become overconfident? (Score:4, Insightful)
First, economics has very little models that are actually worth a damn. They cling to unworkable ideas like the utility-maximizing individual as a base model of human behaviour, doing their best to avoid the problem that neither have they any actual definition of utility, nor is there such a maximizing individual anywhere to be found in the wild.
I think most of the failure of the Economics field as a science already stems from that. The "Homo Economicus" is a pure fantasy construct. Most people are not even capable to see the factors that would allow them to maximize their personal gain. And most people would not do it because of ideological reasons, political reasons, wold-view, etc. even if they could. This does include negatives (racism, nationalism, religion, etc.) and positives (altruism or wanting a future for the human race). Hence all this "science" is built on a model that does, maybe, apply to 5% of the population, maybe less.
That will not work and cannot work.
As a side note, I had lecture about this stuff (as an elective during my PhD) by a sociologist. They do the math a bit worse than the economists, but the entertaining and learning factor is very high as sociologists have some real understanding of the world and of people, and they are all cynics. Hence if you have a choice, select a lecturer from sociology.
Re: (Score:2)
There is a certain utility maximising moment that might happen in one's life. If you have your last fiver in your pocket, and need to buy food for an unknown period of time. An economist would say that one should look for the cheapest nutrients per unit weight available and go all in. Turns out many people go for a big mac or such - who knows when the next cash is going to arrive, might as well get something to cheer one up.
I have done both. Cannot say one approach is better than the other.
But I'll take a s
Re: (Score:2)
I strongly suspect that your 3rd point is STRONGLY related to your 2nd. The people behind the second point REALLY don't want the rest of us to actually look at GDP vs. wages. Or profits vs. wages. They also seem to want to avoid discussion of retail price vs. marginal cost of production.
Your 1st point is also valid. I would add that many sellers do their best to make sure that even if that mythical beast existed, he/she wouldn't have enough reliable information to make a rational decision.
Re: (Score:2)
Lack of information is definitely a major showstopper. Any such a mythical utility maximizing creature presupposes total information availability, and endless time, energy and brain for the creature to figure everything out. With this also goes down the drain the workability of e.g. libertarianism. Every single decision in such a framework becomes a full time job.
Sadly, or amusingly, democracy itself also has the same presuppositions on the base level. This requirement can be worked around if the individual
Re: (Score:2)
The interesting thing here is they're also under-confident in the short term.
I wonder if this is them struggling to correctly predict the probability of unusual events. For instance, they consider there's a 10% chance something screwy will happen to throw off their estimate. But they use roughly the same 10% estimate in the short term as the long term, not allowing for the fact that it increases (kinda) linearly with time.
Re: (Score:2)
Not all humans are. Being overconfident is sort-of a group activity though and requires you to ignore parts of reality, something for which "groupthink is great! The "independent thinkers" will usually not be overconfident, but these are only something like 10-15% of the population. And the rest hates it when these people make forecasts because that results in things like "climate is changing and it will get bad unless we really restrict ourselves now" or "get vaccinated to slow down the pandemic" or "Micro
Re: (Score:2)
As for your political angle, this is actually a brilliant way of having a scapegoat for all the missteps and failed promises. 'It wasn't us it was him'.
People also react to economic forecasts (Score:1)
Which can change their outcome.
Forecasters? Please! (Score:1)
These people are more like astrologers, or phrenologists, phony as a three dollar bill. Actually they are just mouthpieces for the people that pay them, Goldman Sachs, Chase, Wells Fargo, all the same people that bring on your regular market collapses.... All accuracy depends on the amount of inside information they have.
Obviously (Score:2)
Those that are not will leave their job and look for a better life. Also, anybody that wants people to invest in something will always only hire the overconfident forecasters. Next question?
Humans in general are overconfident (Score:4, Insightful)
75% of drivers say they are above average drivers.
Software engineers (and everybody) routinely underestimates how long a project will take.
Huge numbers of people play the lottery routinely.
It's a thing, people are unreasonably optimistic.
Re: (Score:2)
75% of drivers say they are above average drivers.
Software engineers (and everybody) routinely underestimates how long a project will take.
Huge numbers of people play the lottery routinely.
It's a thing, people are unreasonably optimistic.
How are you so confident about this?
Care of give references? Is that 75% number made up. How likely is it that exactly 75% say something. How do you define an average driver. Who are the people who think they are less than an average driver? What metrics did they base their evaluation on? Where was the survey conducted? How many people participated in the survey? What was the method of answering the survey? What was the distribution of the people in the survey in gender, age, profession, income, wealth etc?
Re: (Score:2)
Here's a link regarding the 75% of drivers number: https://www.linkedin.com/pulse... [linkedin.com]
I didn't say "everybody" underestimates project size, I said engineers _routinely_ do this. Words matter.
Do you also dispute that huge numbers of people play the lottery? Would you like a source?
You seem really worked up about this!
Re: (Score:2)
Here's a link regarding the 75% of drivers number: https://www.linkedin.com/pulse... [linkedin.com]
I didn't say "everybody" underestimates project size, I said engineers _routinely_ do this. Words matter.
Do you also dispute that huge numbers of people play the lottery? Would you like a source?
You seem really worked up about this!
Why are you so confident that your 75% think blah blah is true and accurate. Your over-confidence about this is disconcerting.
This is the abstract from your reference's reference. The paper says something completely different.
"We examined whether people recognized that others might disagree with their high self-assessments of driving ability, and, if so, why. Participants in four experiments expressed a belief that others would assess them as worse drivers than they assessed themselves. This difference appe
Re: (Score:2)
Apparently, my own overconfidence in this study is in line with my original statement that humans are overconfident.
I have no idea what your reasoning is, I don't follow your logic at all, so I have no answer.
The dunning-kruger effect in effect. (Score:2)
When it comes to professional forecasters, it's their job to be overconfident because they're selling bullshit. You've got to believe your own lie to sell it.
Incentives. (Score:3)
Not seeing the forest for the trees (Score:2)
Analysts being right in the short-term doesn't surprise me. They're focused on the finance machinery and the minutiae of its operations. As long as everything within that system is functioning as intended, they have a good grasp of what the different levers and indicators within that system are doing.
When the economic system starts running into the hard limits of external reality, finance's tools start becoming less relevant. They quit working as intended or understood within the bubble of finance. This is
I want to ... (Score:2)
Duh. (Score:3)
Who do you think sells more units: a forecaster with confidence, or a forecaster with doubts?
FRB (Score:2)
Most likely, the FRB is overly confident in its study.