Donating effectively is a tough problem. Finding highly effective organisations is a time-consuming, resource-intensive job. Large foundations (such as Open Philanthropy or the Gates Foundation) spend a significant amount of time researching these questions, but for individual donors, making these decisions can be hard. Furthermore, given the size of any individual donor’s donation, it’s often hard for each person to justify investing a lot of time into investigating the most effective use of their money.
For these reasons we often recommend donors give to expert-led grantmaking funds or follow the advice of trusted charity evaluators. But choosing which funds or evaluators to rely on is also a difficult task, and it may be the case that there is no relevant expert who you believe captures all the considerations.
Another option we recommend for some donors is to participate in a donor lottery. Donor lotteries give donors a stake in directing a larger pot of money than they’d typically decide on just by themselves. Each donor’s probability of winning the lottery is proportional to their donation, and the winner can spend time thinking and researching their decision and then recommend grants for the whole pot. In expectation, each donor is granting the same amount of money to their preferred charities as they would have if they had donated directly. However, for the donor that wins, the larger pot of money makes it worthwhile to spend more time and energy researching where the money should go. It also creates an economy of scale for the other individual donors in the lottery, as only one person is required to do the research, and gets around situations where a donor would like to give to an organisation with a minimum donation size.
For more information on the genesis of the donor lottery, we recommend reading Carl Shulman’s post introducing the idea of donor lotteries or EA Fund’s post explaining why contributing to a donor lottery might be an especially impactful way to give.
There are three significant dates in the Giving What We Can Donor Lottery:
Learn more about how previous donor lotteries (run by us or by others) have been distributed by reading these donor lottery reports:
If you have any questions please get in touch.
Donors enter the lottery by making a donation to it. The odds of winning are given by the size of the donation relative to the lottery’s block size.
For example, if the block size is $100,000, and you enter the lottery with a $2,000 donation, your probability of winning will be 2%:
2,000 / 100,000 = 0.02
Winners will be selected by taking the first 10 hexadecimal digits of the value of the NIST Randomness Beacon at the lottery’s draw date. The entrant whose ‘ticket’ (see below) contains the winning number will be declared the winner. If there is more than one block, all blocks will use the same winning number. The NIST Beacon provides a publicly-verifiable source of randomness. Links to each block’s NIST beacon entry will be posted on the Lock Date.
Taking the first 10 hexadecimal digits means winning lottery draws will be in the range [0, (16^10 - 1)].
For example, let’s say a lottery's draw date is at midday on December 1, 2017 (UTC). The NIST Beacon value for this time (given by the Unix Timestamp 1512129600) was:
02FDF3FCCE145B93C3FA46DFDDA6113C770006CBEC54C6D7ACF21B005CAC3B91FAF9B0C4ED74A070B70B19D740CEDAEE20D6C1B96D59006E596DEB94C0266888
Taking the first ten hexadecimal digits gives us 02FDF3FCCE (or 12850560206 in decimal).
Entrants will be assigned a ‘ticket’, which is an interval in the range [0, (16^10-1)].
The size of the interval is given by:
⌊(Probability of winning) * (16^10 - 1)⌋ where ⌊n⌋ = floor(n)
The actual start and end values of a given entrant’s ticket interval will be determined by distributing these intervals across the entire range.
For example, let’s say there are three entrants into a lottery with a block size of $100,000:
The entrants’ ticket sizes are as follows: ``` Range 'R' = (16^10 - 1)
= 1099511627775
Alice ticket size =⌊(20,000 / 100,000) * R⌋
= 219902325555
Bob ticket size = ⌊(10,000 / 100,000) * R⌋
= 109951162777
Carol ticket size = ⌊(30,000 / 100,000) * R⌋
= 329853488332
```
Tickets are allocated by dividing up the entire range of possible winning numbers:
Alice = [ 0, 219902325554] Bob = [219902325555, 329853488331] Carol = [329853488332, 659706976664] Unallocated = [659706976665, 1099511627775] (lottery guarantor)
Represented visually, this looks like (G is the lottery guarantor):
Taking the winning number from above, 12850560206, we declare Alice the winner, as 12850560206 falls within her interval [0, 219902325554].
If the lottery takes more in donations than the block size, we split the range into multiple blocks. Multiple blocks ensure that there is no cap on the number of donors who may enter the lottery, while ensuring that the guarantor's liability is capped at the block size.
Let's say that a fourth donor, Dave, enters the lottery before it closes, with a $50,000 donation. As this will overflow the block, a second block is now in play. Dave now has a ticket range that covers $40,000 in the first block, and an additional $10,000 in the second. Note that Dave's chance of winning is still 50%, and as there is only one winning number, he can only win in one of the blocks.
At draw time, winners are selected in each block. Using the same winning number as in the first example, Alice wins the first block, and Dave wins the second. They each may independently recommend $100,000 worth of grants.
Ticket intervals will be assigned to donors at the lock date of the lottery.