Suppose we are in November 2020 and you are a climate-focused donor who plans to donate $1 million to the Founders Pledge Climate Change Fund (CCF).
If Biden wins the election then, based on your research, you expect the effectiveness of your donation will rise by about 10x. Suppose the baseline cost-effectiveness of the CCF is roughly $1/tCO2e (tonne of CO2 equivalent), so under Biden you expect it to be better at $0.1/tCO2e. Also, suppose that in either case, Biden or Trump, you see ‘investing to give’ as less effective than donating now (at least with this money you have earmarked for giving now).
You believe Biden has a 70% chance of winning (see NYT article). You see that on Betfair you can get 1.5:1 odds on Biden.
If you just donate, then your donation averts 7.3 million tCO2e in expectation (see below the main post for the calculations).
But if you bet $1m on Biden, with the commitment to donate the potential $1.5m win, then your expected impact is to avert 10.5 million tCO2e.
So, because almost all the impact you expect from your donation occurs when Biden wins, you can increase your expected impact by more than 40%.
This is an example of event-driven mission hedging. That is, investing around an event in a way such that your returns are correlated with the effectiveness of how you intend to use them.
If you already thought that donating to CCF was the best use of your money at the time, then with this mission hedging strategy you would think it is even better. It is also possible that if you didn’t think CCF was your best donation option before, then with this strategy it might become your best option—see ‘More general context’ below for a discussion of this.
Many people might not feel comfortable using this mission hedging strategy. If you lean in this direction, I’d welcome it if you share your reasoning. I’d encourage you to not simply disagree with the numbers—they are plausible and documented, but ultimately only illustrative. Rather, conditional on having these beliefs at the time, are there other factors that would cause you to avoid this strategy?
For others, the fact that this strategy is emotionally unattractive to many, and hence neglected, makes it all the more attractive to consider. Hauke Hillebrandt wrote a comprehensive introduction to mission hedging several years ago. Yet practical uptake and application seems to have been slow, except perhaps among some AI investors. A minor indicator of this is that there was no EA Forum tag for ‘mission hedging’ until I added one for this post. Further exploration of this concept seems worthwhile, whether for short-horizon strategies like in this post or long-horizon ones like ‘buy AI stocks’.
Prediction markets are helpful in the case I’ve presented here because they offer a direct trade on an event which is expected to dramatically affect CCF’s effectiveness. I’ve talked about $1 million here just to emphasize that this strategy could have been applied at this scale. Traditional financial markets could be used to apply this strategy more frequently and at greater scale, but with less clear correlations between financial returns and effectiveness. Just like with regular event-driven trading this strategy requires expertise. Altruistic investors could hire traders to spot events that are likely to influence effectiveness in their priority cause areas and make corresponding trades.
Would you have tried this strategy last November? Did you?
Acknowledgements
Thanks to Hauke Hillebrandt and Johannes Ackva for helpful comments on earlier drafts. This of course does not imply their endorsement of all aspects of this post.
Additional notes
Main calculations
The expectation for just donating is
$1 million∗(70%/($0.1/tCO2e)+30%/($1/tCO2e))=7.3 million tCO2e
The expectation for mission hedging is
$1 million∗1.5∗70%/($0.1/tCO2e)=10.5 million tCO2e
More scenario analysis
If your probability for a Biden win was 90% the expected return on the bet, with the same odds, rises to 48%. It falls to 36% for a win probability of 50% - this shows there is value in mission hedging even if the odds on offer are worse than what you would consider fair.
Similarly, sticking with a 70% probability as a reference point, the expected ‘return’ on the mission hedging bet falls to 28% if you could only get 1.33:1 odds on Biden. But, according to the New York Times, there were times during the election when the odds were as much as 7:1. With those odds the expected mission hedging return would be 571%.
More general context
Another consideration is that while a second Trump term may make US-focused advocacy less effective, it could make donating to climate change efforts globally more valuable[1]. Or, if global climate progress is made less tractable under Trump, then it may be most effective to simple switch cause area in this scenario. Or only switch into climate if Biden wins.
Suppose that normally you are a consistent donor to a GiveWell top charity (GWC). However, you have looked into CCF and while you believe that under another Trump term it will only be 0.2x as effective as the GWC, in the scenario of Biden’s first year it will be 2x GWC[2].
If you wait for Biden to win and donate to CCF, and otherwise continue with your GWC donation, the expected value is equivalent to 1.7 million dollars to the GWC.
If you do the mission hedging bet (with the original odds and probability from above) then the expected value is equivalent to 2.1 million dollars to GWC.
So there is still a 24% expected ‘return’ to mission hedging in this more general context.
General formula
p=your probability for the event that leads to high effectivenessq=multiplier for increased effectiveness, relative to baseline of 1pb=probability equivalent to the oddsc=transaction costs as a percentage of amount bet
Naturally MH will be higher for greater q and smaller c and pb. Additionally betting on unlikely events with low p (and low pb to match) seems most valuable. For example, if Biden was a 10:1 underdog and you agreed so p=0.1, then MH=5.26.
What events-based mission hedging requires
Ability to forecast how the outcome of an event (election, court case, …) will change the cost-effectiveness of donations to a particular organization or cause area
For one outcome of the event to lead the target organization to have higher cost-effectiveness than alternative donation opportunities
A reasonably liquid investment whose financial returns are highly correlated with the event
A back of the envelope way to actually arrive at these multipliers as a rough but genuine assessment is to take the Mortality Cost of Carbon of 2 x 10-4 excess deaths / tCO2e from Bressler (2020) and multiply it by the $1 / tCO2e cost-effectiveness as a baseline for CCF that I used in the main text. Then at baseline level CCF would seem to have cost-effectiveness of $5,000 / life saved. GiveWell’s most recent explicit cost per life saved estimates for specific projects of their top charities range from $783/life to $9207/life.
Event-driven mission correlated investing and the 2020 US election
Suppose we are in November 2020 and you are a climate-focused donor who plans to donate $1 million to the Founders Pledge Climate Change Fund (CCF).
If Biden wins the election then, based on your research, you expect the effectiveness of your donation will rise by about 10x. Suppose the baseline cost-effectiveness of the CCF is roughly $1/tCO2e (tonne of CO2 equivalent), so under Biden you expect it to be better at $0.1/tCO2e. Also, suppose that in either case, Biden or Trump, you see ‘investing to give’ as less effective than donating now (at least with this money you have earmarked for giving now).
You believe Biden has a 70% chance of winning (see NYT article). You see that on Betfair you can get 1.5:1 odds on Biden.
If you just donate, then your donation averts 7.3 million tCO2e in expectation (see below the main post for the calculations).
But if you bet $1m on Biden, with the commitment to donate the potential $1.5m win, then your expected impact is to avert 10.5 million tCO2e.
So, because almost all the impact you expect from your donation occurs when Biden wins, you can increase your expected impact by more than 40%.
This is an example of event-driven mission hedging. That is, investing around an event in a way such that your returns are correlated with the effectiveness of how you intend to use them.
If you already thought that donating to CCF was the best use of your money at the time, then with this mission hedging strategy you would think it is even better. It is also possible that if you didn’t think CCF was your best donation option before, then with this strategy it might become your best option—see ‘More general context’ below for a discussion of this.
Many people might not feel comfortable using this mission hedging strategy. If you lean in this direction, I’d welcome it if you share your reasoning. I’d encourage you to not simply disagree with the numbers—they are plausible and documented, but ultimately only illustrative. Rather, conditional on having these beliefs at the time, are there other factors that would cause you to avoid this strategy?
For others, the fact that this strategy is emotionally unattractive to many, and hence neglected, makes it all the more attractive to consider. Hauke Hillebrandt wrote a comprehensive introduction to mission hedging several years ago. Yet practical uptake and application seems to have been slow, except perhaps among some AI investors. A minor indicator of this is that there was no EA Forum tag for ‘mission hedging’ until I added one for this post. Further exploration of this concept seems worthwhile, whether for short-horizon strategies like in this post or long-horizon ones like ‘buy AI stocks’.
Prediction markets are helpful in the case I’ve presented here because they offer a direct trade on an event which is expected to dramatically affect CCF’s effectiveness. I’ve talked about $1 million here just to emphasize that this strategy could have been applied at this scale. Traditional financial markets could be used to apply this strategy more frequently and at greater scale, but with less clear correlations between financial returns and effectiveness. Just like with regular event-driven trading this strategy requires expertise. Altruistic investors could hire traders to spot events that are likely to influence effectiveness in their priority cause areas and make corresponding trades.
Would you have tried this strategy last November? Did you?
Acknowledgements
Thanks to Hauke Hillebrandt and Johannes Ackva for helpful comments on earlier drafts. This of course does not imply their endorsement of all aspects of this post.
Additional notes
Main calculations
The expectation for just donating is
$1 million∗(70%/($0.1/tCO2e)+30%/($1/tCO2e))=7.3 million tCO2e
The expectation for mission hedging is
$1 million∗1.5∗70%/($0.1/tCO2e)=10.5 million tCO2e
More scenario analysis
If your probability for a Biden win was 90% the expected return on the bet, with the same odds, rises to 48%. It falls to 36% for a win probability of 50% - this shows there is value in mission hedging even if the odds on offer are worse than what you would consider fair.
Similarly, sticking with a 70% probability as a reference point, the expected ‘return’ on the mission hedging bet falls to 28% if you could only get 1.33:1 odds on Biden. But, according to the New York Times, there were times during the election when the odds were as much as 7:1. With those odds the expected mission hedging return would be 571%.
More general context
Another consideration is that while a second Trump term may make US-focused advocacy less effective, it could make donating to climate change efforts globally more valuable[1]. Or, if global climate progress is made less tractable under Trump, then it may be most effective to simple switch cause area in this scenario. Or only switch into climate if Biden wins.
Suppose that normally you are a consistent donor to a GiveWell top charity (GWC). However, you have looked into CCF and while you believe that under another Trump term it will only be 0.2x as effective as the GWC, in the scenario of Biden’s first year it will be 2x GWC[2].
If you wait for Biden to win and donate to CCF, and otherwise continue with your GWC donation, the expected value is equivalent to 1.7 million dollars to the GWC.
If you do the mission hedging bet (with the original odds and probability from above) then the expected value is equivalent to 2.1 million dollars to GWC.
So there is still a 24% expected ‘return’ to mission hedging in this more general context.
General formula
p=your probability for the event that leads to high effectivenessq=multiplier for increased effectiveness, relative to baseline of 1pb=probability equivalent to the oddsc=transaction costs as a percentage of amount bet
Then
EV of normal donation:VN=p∗q+(1−p)
EV of mission hedged donation:VMH=ppb∗(1−c)∗q
So
Mission Hedging multiplier:MH=VMHVN=ppb∗(1−c)∗q1+p∗(q−1)
It helps to consider some particular cases,
Limit where p*q large:MH≈(1−c)/pb
Limit where p is small:MH≈(p/pb)∗(1−c)∗q
Naturally MH will be higher for greater q and smaller c and pb. Additionally betting on unlikely events with low p (and low pb to match) seems most valuable. For example, if Biden was a 10:1 underdog and you agreed so p=0.1, then MH=5.26.
What events-based mission hedging requires
Ability to forecast how the outcome of an event (election, court case, …) will change the cost-effectiveness of donations to a particular organization or cause area
For one outcome of the event to lead the target organization to have higher cost-effectiveness than alternative donation opportunities
A reasonably liquid investment whose financial returns are highly correlated with the event
Note that CCF is a global fund, so it would indeed most likely switch from a US focus to a global focus in this case.
A back of the envelope way to actually arrive at these multipliers as a rough but genuine assessment is to take the Mortality Cost of Carbon of 2 x 10-4 excess deaths / tCO2e from Bressler (2020) and multiply it by the $1 / tCO2e cost-effectiveness as a baseline for CCF that I used in the main text. Then at baseline level CCF would seem to have cost-effectiveness of $5,000 / life saved. GiveWell’s most recent explicit cost per life saved estimates for specific projects of their top charities range from $783/life to $9207/life.