Non Standard Preferences (Time Inconsistency)-Ashrafe (2006) ”Tying Odysseus to the mast: Evidence from a commitment savings product in the Philippines.” Lecture/ Class 3 Flashcards
What was the paper , ”Tying Odysseus to the mast:
Evidence from a commitment savings product in the Philippines.” by Ashraf About?
It wanted to test to see if those who had time- inconsistent preferences would be aware of this (not naive) and so would get committment devices and save more
What are time- inconsistent preferences
Maybe people in this bank were less naive?
When you are impatient when considering near-future outcomes but more patient for outcomes in the distant future.
This is because their discount factor δ is not constant (violating standard theory) it is large in the near future and lower in the far future
What are hyperbolic preferences
When you’re willing to wait a long time for a high
amount but not willing to wait a short time for a small amount (time -inconsistent)
δ decreases over time (not constant violating standard theory)
What was the methodology?
This was a field experiment where they randomnly (RCT) chose 1777 households that were currently or previously part of a specific bank and asked them questions to find out if they had hyperbolic preferences.
Then half of these people were put in the treatment group and offered a committment savings product called SEED that doesn’t allow the owner to access their money util a certain amount of time has passed or if they’ve reached a savings goal. it has no other benefits.
the other half were a control group and not offered SEED
What were the two ways that people could deposit money into SEED?
They could keep a box (like a piggy bank) that they could put small amounts of money into and then bring this to the bank to deposit (saves them from daily trips to the bank). Although this box could only be unlocked by the bank, it wouldn’t be too hard for the saver to break the box to get the money out.
or they could automatically have transfers from a checking account to SEED
What were the two main areas of interest for estimation?
Finding out how many from the treatment group (those offered SEED) opted into SEED
How much money people saved
What were the two things that Ashraf was trying to measure that may influence how much a person saves
How people with hyperbolic preferences (time-inconsistent) save
How impatient people save (if they’re aware of their impatience they’ll want a commitment advice and opt into SEED)
People deemed impatient if they’ll take small immediate rewards but reject big rewards that are delayed
What was the aim of the survey and give an example of a question asked?
All people both T and C were questioned to see if they had time-inconsistent preferences.
Would you prefer 200 pesos today or 300 pesos in one month?
Would you prefer 600 pesos in 6 months or 700 pesos in 7 months?
These questions are the same in that the option is one month and 100 pesos apart so someone who is impatient and chooses the immediate reward would be expected to pick the immediate in both of the options. Choosing the immediate reward in the near term frame but not the distant term shows time-inconsistency and is hyperbolic going against the standard model.
Were time inconsistent people sophisticated (aware of their hyperbolic prefrences and so took SEED)
yes many were, The biggest predictor of take-up of the commitment savings was hyperbolic preferences, the effect was strong for women but negligible for men.
What could explain why the effect of men having hyperbolic prefrences on their likelihood to take up SEED was negligible?
The reason for the statistically INsignificant effect on men, may be that due to women managing the finances in most households in the Philippines, men may see less reason for them to take-up SEED as they have a smaller role in household finances.
What are some criticisms you could make on the paper?
Only about half the people that they contacted agreed to do the survey, could it of been a certain type of person that agreed to do the survey, was it those that were more committed and likely to save anyways?
That people with SEED were able to either to be able to withdraw their money once they reached a certain savings goal or once a certain date had passed could create bias. Maybe those that chose to set a certain savings goal were more sophisticated and aware of their impatience so this could create an overestimation of results.
The researchers should have randomly assigned people to be able to withdraw past a date or past a goal was met.
How does the standard theory allow for discounting?
It allows us to prefer to have goods today and we discount tomorrow by δ however this δ doesn’t change (is constant)
i.e. how much you discount getting chocolate on wednesday rather than tuesday is the same amount by which you discount getting it on thursday rather than wednesday.
Is the discount rate the same as the discount factor?
No
discount factor is denoted δ
What is hyperbolic discounting?
When the discount factor falls over time rather than remaining constant
How do we amend the utility function shown by standard theory to allow for quasi hyperbolic discounting?
Include a unique discount factor B
Like Regular Ut function but with a B next to the δ
lecture 3 slide 18
