Catherine from GiveWell here. Thanks for your post—it’s generated some good discussion!
First, I want to confirm a few aspects of our banking that have been discussed here:
A large proportion of the funding that we hold at any point in time is for making grants to our recommended charities. We typically grant those funds within 2-5 months of receiving them. Therefore, our balance fluctuates throughout the year as funds are received and granted.
The vast majority of donations to GiveWell are made in December, which means that both our end-of-year and beginning-of-year balances are significantly higher than the typical balance on any given day of the year. Our January balance is higher because we have not yet granted the December donations, and because some year-end donations are processed in January, such as checks (which tend to be larger donations) that are mailed to us on December 31.
Beyond funding for granting, we also maintain a balance of operating expenses.
We agree that consideration of where to hold and invest funds is important. Last year, we began work to revisit our banking practices. We also adopted an investment policy in December 2020 that enables us to invest a portion of our available cash balances in financial vehicles, with the following goals: preserving capital, meeting our liquidity requirements, minimizing volatility, and maximizing our after-tax return.
We are continuing our project to revisit our banking this year. In addition to the interest rates offered on accounts, our decisions about where to hold funds will also factor in our additional needs, such as strong customer service, security, understanding of non-profit vs. for-profit needs, user interface, connection with our various tools and platforms, and international operability. We plan to take this work forward in 2021.
Thanks for your response Catherine! It’s great to hear that GiveWell is moving in this direction!
Makes sense regarding those three aspects of GiveWell’s banking, and completely agree that our estimation methodology is likely in GiveWell’s case to overestimate its average cash balance during the year. For the benefit of readers, as mentioned in our article, whether the estimate is accurate, too low, or too high will vary by charity and there is no way to increase the accuracy for an estimation methodology that works across charities due to the limitations of the Form 990 data.
Other commenters have mentioned that the federal funds rate is now close to zero. This article was primarily written before COVID-19 and the federal funds rate seemed like an appropriate and conservative benchmark at the time. I did not update the benchmark before the article was published because I received feedback from reviewers that the magnitude of the impact was large regardless of the exact interest rates. This is a sentiment that I agree with.
If the exact rate is literally 0% for the next few years (I did not intend for the federal funds rate to be used as a forward-looking estimate) that would indicate the benefit to charities is pretty low. I think a better proxy for the interest rate charities could have earned in the past and can earn now and into the future is Ally Bank’s high-yield savings account rate. Ally Bank is a large institution and has over 10 years of historical interest rate data available. Their rate has historically been close to high-yield savings rates offered by most banks, and just a touch lower than the best options.
At this exact point in time, Ally Bank’s rate is 0.5%. Here is the link to a historical rate chart with data since 2009: https://www.depositaccounts.com/banks/ally-bank.html#a305301. Ideally, GiveWell and other charities earning interest on U.S. dollar deposits should choose banking options that have historically earned and currently earn something close to this rate. Unfortunately, I do not think organizations can open accounts at Ally Bank at this time because they only cater to individuals, but there are many options for organizations out there that are competitive with Ally Bank’s rate. We’re happy to provide suggestions to GiveWell or any other organization.
As you can see from the chart, back when the federal funds rate was close to 0% after the 2008 recession, Ally Bank’s interest rate hovered around 1%. That was one of the reasons my counterfactual impact estimate for GiveWell used a 1% rate. In retrospect, I should have elaborated on how I generated that estimate in the article itself.
I understand that factors like customer service, security, technology integrations, etc influence GiveWell’s decision making around its banking needs. I have two thoughts on the matter.
Firstly, a high-yield savings account simply exists to maximize the yield on a nonprofit’s cash, whereas a checking account accommodates most of an organization’s banking needs. A checking account does not need to be selected on the basis of its interest rate, whereas this is much more important for a savings account, where an nonprofit’s appropriate savings balance (e.g. most of its cash that isn’t immediately needed) is stored.
Secondly, GiveWell and other charities can also select the option of having an account at a major brokerage firm such as Vanguard or Fidelity which may be more trustworthy and integrated with more options than a lesser know savings account provider (although FDIC insurance means that most regulated U.S. banking options are safe). Charities can hold these funds inside a money market account (offered by most brokerage providers as well as a lot of third parties), which is designed to not change in value, or another low-risk option. My article from last year contains more information on different options for charities, including analyzing the risk/security of both bank accounts and money market funds which you mentioned GiveWell is considering, as well as concrete guidance on selecting a brokerage account and the fund(s) to hold within it.
Hi Brendon and Tharun,
Catherine from GiveWell here. Thanks for your post—it’s generated some good discussion!
First, I want to confirm a few aspects of our banking that have been discussed here:
A large proportion of the funding that we hold at any point in time is for making grants to our recommended charities. We typically grant those funds within 2-5 months of receiving them. Therefore, our balance fluctuates throughout the year as funds are received and granted.
The vast majority of donations to GiveWell are made in December, which means that both our end-of-year and beginning-of-year balances are significantly higher than the typical balance on any given day of the year. Our January balance is higher because we have not yet granted the December donations, and because some year-end donations are processed in January, such as checks (which tend to be larger donations) that are mailed to us on December 31.
Beyond funding for granting, we also maintain a balance of operating expenses.
We agree that consideration of where to hold and invest funds is important. Last year, we began work to revisit our banking practices. We also adopted an investment policy in December 2020 that enables us to invest a portion of our available cash balances in financial vehicles, with the following goals: preserving capital, meeting our liquidity requirements, minimizing volatility, and maximizing our after-tax return.
We are continuing our project to revisit our banking this year. In addition to the interest rates offered on accounts, our decisions about where to hold funds will also factor in our additional needs, such as strong customer service, security, understanding of non-profit vs. for-profit needs, user interface, connection with our various tools and platforms, and international operability. We plan to take this work forward in 2021.
Thanks for your response Catherine! It’s great to hear that GiveWell is moving in this direction!
Makes sense regarding those three aspects of GiveWell’s banking, and completely agree that our estimation methodology is likely in GiveWell’s case to overestimate its average cash balance during the year. For the benefit of readers, as mentioned in our article, whether the estimate is accurate, too low, or too high will vary by charity and there is no way to increase the accuracy for an estimation methodology that works across charities due to the limitations of the Form 990 data.
Other commenters have mentioned that the federal funds rate is now close to zero. This article was primarily written before COVID-19 and the federal funds rate seemed like an appropriate and conservative benchmark at the time. I did not update the benchmark before the article was published because I received feedback from reviewers that the magnitude of the impact was large regardless of the exact interest rates. This is a sentiment that I agree with.
If the exact rate is literally 0% for the next few years (I did not intend for the federal funds rate to be used as a forward-looking estimate) that would indicate the benefit to charities is pretty low. I think a better proxy for the interest rate charities could have earned in the past and can earn now and into the future is Ally Bank’s high-yield savings account rate. Ally Bank is a large institution and has over 10 years of historical interest rate data available. Their rate has historically been close to high-yield savings rates offered by most banks, and just a touch lower than the best options.
At this exact point in time, Ally Bank’s rate is 0.5%. Here is the link to a historical rate chart with data since 2009: https://www.depositaccounts.com/banks/ally-bank.html#a305301. Ideally, GiveWell and other charities earning interest on U.S. dollar deposits should choose banking options that have historically earned and currently earn something close to this rate. Unfortunately, I do not think organizations can open accounts at Ally Bank at this time because they only cater to individuals, but there are many options for organizations out there that are competitive with Ally Bank’s rate. We’re happy to provide suggestions to GiveWell or any other organization.
As you can see from the chart, back when the federal funds rate was close to 0% after the 2008 recession, Ally Bank’s interest rate hovered around 1%. That was one of the reasons my counterfactual impact estimate for GiveWell used a 1% rate. In retrospect, I should have elaborated on how I generated that estimate in the article itself.
I understand that factors like customer service, security, technology integrations, etc influence GiveWell’s decision making around its banking needs. I have two thoughts on the matter.
Firstly, a high-yield savings account simply exists to maximize the yield on a nonprofit’s cash, whereas a checking account accommodates most of an organization’s banking needs. A checking account does not need to be selected on the basis of its interest rate, whereas this is much more important for a savings account, where an nonprofit’s appropriate savings balance (e.g. most of its cash that isn’t immediately needed) is stored.
Secondly, GiveWell and other charities can also select the option of having an account at a major brokerage firm such as Vanguard or Fidelity which may be more trustworthy and integrated with more options than a lesser know savings account provider (although FDIC insurance means that most regulated U.S. banking options are safe). Charities can hold these funds inside a money market account (offered by most brokerage providers as well as a lot of third parties), which is designed to not change in value, or another low-risk option. My article from last year contains more information on different options for charities, including analyzing the risk/security of both bank accounts and money market funds which you mentioned GiveWell is considering, as well as concrete guidance on selecting a brokerage account and the fund(s) to hold within it.