This is the best solution video that I have watched on procrastination. I finally have a solution for a lifetime, which is now or never. Thank you so much!
The great thing about your videos is the balance you strike between “real” Econ and layman’s, “real” life applications. There’s a great book here in the making. And also IMHO with a bit of additional investment into production this channel would be really really phenomenal and valuable resource for a bigger audience.
Honestly, this has been such a helpful explanation of the model. I knew that hyperbolic discounting could be used in systems like save more tomorrow but I didn't understand how it really works. When I was going through a lot of the academic papers on hyperbolic discounting in finance too I found it really hard to wrap my head around what was going on so this has really helped me set that straight thank you.
Corporate Finance - DCF/Present Value discounting techniques + applied to Behavioral Economics. Totally cool. Myth Optimizing can be added or discounted to the discount rate, which could also include other externalities to calculate future discounted cash flows in the future.
Hi, I am gonna do PhD in Behavioural Economics. Could you please suggest some readings on time discounting and related field? This video was really helpful in clearing my basics. Thanks.
Clears up a lot of haze on the topic ;). One hazy item left for me: Was r in the Excel tips example calculated or assigned? If calculated, what's that ROI calculation?
Send me an email explaining. I can't guarantee I'll have time for it, but I'm interested to hear about your community, and I'll help if it's a smaller task/model.
Newbie here. For this model, the assumption is that there is effort from the individual, right? Whereas, when the task is to receive something, instant gratification has more value than delayed gratification. Did I understand this correctly ?
I procrastinate IF I have something, which saves me ~5 minutes, i.e., doing dishes. Then instead of 0, I have 5 + 76, i.e., ~81 minutes from procrastinating. Then procrastination would make sense in the classic utilitarian model too! That my contention with utilitarian models, one minor tweak here and there, and it makes sense again.. how do we do it's Beta Delta, not the positive value from doing something else for the 120 minutes
Hi Ashley, thanks so much for your explanation. I have a question regarding low-/high discounting rate. At my uni my teacher said: "quasi hyperbolic discounting had a high-short-term discount rate and a lower long-term discount rate." However for me it would only make sense the other way around: short term -> low discount rate, long term -> high discount rate. Am I wrong? I’d appreciate your opinion/explanation on this. Best regards
In the short term, discount rate includes both beta and delta/r together, so you are discounting more heavily in the short run. The long-run discount rate is less important, so you are discounting less between any two periods far into the future. The confusion between you and your teacher may be in the way we refer to discounting, which is a little wonky. High levels of discounting = high r = low delta. When people use the term "discount rate", that generally refers to r, while discount factor is 1/(1+r)=delta. I therefore suspect the difference in opinion between you and your professor is about language. High discount rate = low discount factor. Your professor is correct, technically using "discount rate". You would be correct if you switched to "discount factor".
This is the most abstruse way of analyzing the behavior of f(r) = 100 b (1+1/r) that i've ever seen 99% of economists should really consider a remedial class in algebra and analysis.
It must look like "the most abstruse way of analyzing the behavior of f(r) = 100 b (1+1/r)" to you, mainly because that is not at all what this video is about. It looks like you tried to simplify the overall utility function U(r) given in the video: U(r) = 100 + 100*b*(1/(1-r)) + 100*b*(1/(1-r))^2 + 100*b*(1/(1-r))^3 + ... The correct way to do that would be by putting 100*b outside the brackets which gives something like: U(r) = 100 + 100*b*[ (1/(1-r)) + (1/(1-r))^2 + (1/(1-r))^3 + ... ] There are several mistakes with f(r) = 100 b (1+1/r): ------------------------------------------------------------------------------------------------------------------------------------------------------------------ 1. The expression (1+1/r) is wrong. It would have to be (1/(1-r)). 2. b can't be put outside the brackets like that. 3. The exponents are missing entirely. 4. Because the exponents are not considered, the entire sum is falsly simplified to just one fraction ------------------------------------------------------------------------------------------------------------------------------------------------------------------ But even if your math would be correct, you are still missing the point completely, because the video is not at all about the behavior of the utility function U(r) with respect to r. In the entire video, r is an exogenously given parameter and never investigated as a variable. Instead, the graphs shown in the video are of the discount functions and their behavior is investigated with respect to time t. You also seem to misunderstand the entire purpose of the model. Discounting models are not about the behavior of some function, but try to describe how people evaluate payoffs and costs which are spread over time. So the question is not: "How does the overall utility function behave with regard to some variable?", but rather: "Which of the given vectors of payoffs and costs maximizes the overall utility function and which discount function is better to explain empirically observed behavior?" The entire comment is just severly confused about both the math and the economics.
You can estimate beta based on people's behavior. For example, if you repeatedly measure their choices related to some time-based decision, you can see what beta fits the data the best. I think this will eventually be used to model behavior in online spaces, where there is a lot of data coming in.
![CHILE vs. BRASIL [1-2] | RESUMEN | ELIMINATORIAS SUDAMERICANAS | FECHA 9](http://i.ytimg.com/vi/5ObJt_71AYc/mqdefault.jpg)
This is the best solution video that I have watched on procrastination. I finally have a solution for a lifetime, which is now or never. Thank you so much!
The great thing about your videos is the balance you strike between “real” Econ and layman’s, “real” life applications.
There’s a great book here in the making. And also IMHO with a bit of additional investment into production this channel would be really really phenomenal and valuable resource for a bigger audience.
Honestly, this has been such a helpful explanation of the model.
I knew that hyperbolic discounting could be used in systems like save more tomorrow but I didn't understand how it really works. When I was going through a lot of the academic papers on hyperbolic discounting in finance too I found it really hard to wrap my head around what was going on so this has really helped me set that straight thank you.
Very good video. Clear, motivated and well explained. I love the emotion with which you speak 💚💚
Corporate Finance - DCF/Present Value discounting techniques + applied to Behavioral Economics. Totally cool. Myth Optimizing can be added or discounted to the discount rate, which could also include other externalities to calculate future discounted cash flows in the future.
beautiful presentation & beautiful presenter.
Hi, I am gonna do PhD in Behavioural Economics. Could you please suggest some readings on time discounting and related field? This video was really helpful in clearing my basics. Thanks.
Just found your channel, I´m stunned by the great content!
Me too.. great material
This is an excellent presentation. Thank you!
Ashely, the math trick that I have difficulty with. Might you explain this in another video? thanks
7:34 its actually a 55% drop. As you explained in the start, the first period drop is 1-beta*delta, not 1-beta
Thanks for the video by the way
Clears up a lot of haze on the topic ;). One hazy item left for me: Was r in the Excel tips example calculated or assigned? If calculated, what's that ROI calculation?
"r" in the "Excel tips" was assigned / exogenous.
This is a great explanation. Thank you!
Awesome presentation Ashley. Have my behavioral economics exam tomo:)
I love behavioural econ. Please do more. I would also like to consult with to help me with a layman's model for my community
Send me an email explaining. I can't guarantee I'll have time for it, but I'm interested to hear about your community, and I'll help if it's a smaller task/model.
This was so interesting and helpful, thanks alot !!
Thank you for your time
Damn that is such a mathematical way of saying 'im lazy'
Newbie here. For this model, the assumption is that there is effort from the individual, right?
Whereas, when the task is to receive something, instant gratification has more value than delayed gratification.
Did I understand this correctly ?
@ashley hodgson.. how do you work out u(Learn tomorrow)= 75..; what mathematic series did you use.. I couldn't figure this out. Thanks!
I procrastinate IF I have something, which saves me ~5 minutes, i.e., doing dishes. Then instead of 0, I have 5 + 76, i.e., ~81 minutes from procrastinating.
Then procrastination would make sense in the classic utilitarian model too!
That my contention with utilitarian models, one minor tweak here and there, and it makes sense again.. how do we do it's Beta Delta, not the positive value from doing something else for the 120 minutes
Hi Ashley, thanks so much for your explanation. I have a question regarding low-/high discounting rate. At my uni my teacher said: "quasi hyperbolic discounting had a high-short-term discount rate and a lower long-term discount rate." However for me it would only make sense the other way around: short term -> low discount rate, long term -> high discount rate. Am I wrong? I’d appreciate your opinion/explanation on this.
Best regards
In the short term, discount rate includes both beta and delta/r together, so you are discounting more heavily in the short run. The long-run discount rate is less important, so you are discounting less between any two periods far into the future. The confusion between you and your teacher may be in the way we refer to discounting, which is a little wonky. High levels of discounting = high r = low delta. When people use the term "discount rate", that generally refers to r, while discount factor is 1/(1+r)=delta. I therefore suspect the difference in opinion between you and your professor is about language. High discount rate = low discount factor. Your professor is correct, technically using "discount rate". You would be correct if you switched to "discount factor".
The decline in utility between period 0 and period 1 for the beta delta model is not 50% but 50%*90% i.e 45%. No?
Yes, that's correct. Beta times delta.
How does this work in regards to procrastinating losing ones virginity, and is it equal for both men and women (and non-binary, LGBTQ, ROFLCOPTER)?
Do you have any spare time for consulting?
thankyou!
its acc so stupid how its explained better for free on RUclips than in the classes i pay £9000 for
Most profs hate their students and would rather spend time on research
We pay for the university name not the teaching
@@hypebeastuchiha9229 on god they hate they students😂, they just expect us to know stuff
YEAH I WOULD FIRE MY ECO PROF FOR U IN EXCHANGE
This is the most abstruse way of analyzing the behavior of f(r) = 100 b (1+1/r) that i've ever seen
99% of economists should really consider a remedial class in algebra and analysis.
It must look like "the most abstruse way of analyzing the behavior of f(r) = 100 b (1+1/r)" to you, mainly because that is not at all what this video is about.
It looks like you tried to simplify the overall utility function U(r) given in the video:
U(r) = 100 + 100*b*(1/(1-r)) + 100*b*(1/(1-r))^2 + 100*b*(1/(1-r))^3 + ...
The correct way to do that would be by putting 100*b outside the brackets which gives something like:
U(r) = 100 + 100*b*[ (1/(1-r)) + (1/(1-r))^2 + (1/(1-r))^3 + ... ]
There are several mistakes with f(r) = 100 b (1+1/r):
------------------------------------------------------------------------------------------------------------------------------------------------------------------
1. The expression (1+1/r) is wrong. It would have to be (1/(1-r)).
2. b can't be put outside the brackets like that.
3. The exponents are missing entirely.
4. Because the exponents are not considered, the entire sum is falsly simplified to just one fraction
------------------------------------------------------------------------------------------------------------------------------------------------------------------
But even if your math would be correct, you are still missing the point completely, because the video is not at all about the behavior of the utility function U(r) with respect to r. In the entire video, r is an exogenously given parameter and never investigated as a variable. Instead, the graphs shown in the video are of the discount functions and their behavior is investigated with respect to time t.
You also seem to misunderstand the entire purpose of the model. Discounting models are not about the behavior of some function, but try to describe how people evaluate payoffs and costs which are spread over time. So the question is not: "How does the overall utility function behave with regard to some variable?", but rather: "Which of the given vectors of payoffs and costs maximizes the overall utility function and which discount function is better to explain empirically observed behavior?"
The entire comment is just severly confused about both the math and the economics.
Yeah I get it but this theory seems useless cause you can never measure the actual number of beta in real situation. Am I wrong? Haha
You can estimate beta based on people's behavior. For example, if you repeatedly measure their choices related to some time-based decision, you can see what beta fits the data the best. I think this will eventually be used to model behavior in online spaces, where there is a lot of data coming in.