Giving USA, the annual compendium on philanthropy published since 1955 has long been the industry standard. With research and development conducted by its partner the Center of Philanthropy at Indiana University it is thorough, complete and has always maintained that it's the best estimate available based on IRS data. The major criticism of course is that Giving USA doesn't come out until six months after the year's close. Though it's not ho-hum it is sort of historical and less effective than a speedier output. We all live more instantaneously (than we probably should) these days; old information simply has less currency. The Holy Grail of information is expert knowledge, accuracy and speed. Giving USA is two out of three.
For the third, speed, there's an attractive newcomer on view. Philanthromax, the brain work of two Texans and, they say, an army of academics has developed a proprietary algorithm set that reports philanthropic data in what is real time in this business: monthly. They've been in business since August 2009; CEO Rob Mitchell told me their model enables them to to have data back to 1969. I hadn't heard of them until an article appeared in Chronicle of Philanthropy a few weeks ago.
I am not an economist. Like most of you I am an avidly interested consumer of philanthropic data. Long ago I had a second undergraduate minor in economics but decades past lost whatever skills I had to analyze data like this. So I got in touch with Philanthromax. I wanted to know how their monthly estimates comported with actual Giving USA data. Their CEO, Rob Mitchell, responded with this chart. I don't have the skills to vet this. I also asked Patrick Rooney for a response and what I got was a "Key Points" .pdf circulated by the Giving Institute's new executive director Geoffrey Brown. It's SOS, all stuff the members have heard before; it doesn't respond in any way to Philanthromax.
Rob Mitchell also said "we are in the process of building sector (religion, health, arts, etc) and source (individual, foundation, corporation, bequest) and geography (state and region). Because of the way the algorithms work, we can build data for the last 4 decades and more importantly, provide a reliable forecast for the coming months. Perhaps the more practical application is that we are able to build Atlas customized forecast algorithms for individual organizations and groups of organizations. This enables very accurate budgeting and now, for the first time, organizations can make decisions about the timing of fundraising campaigns and promotions that will produce better results.”
Professional skeptic that I am everything here is retrospective – not yet real time. Keep a sharp eye on the 2010 estimate!
Here's the (unverifiable by me) data. Giving Institute's Summer Symposium is coming up. I'd like to hear Rob and Patrick discuss their methodology.
| Year | Giving USA | Atlas of Giving | Variance | |||
| 2006 | 295.33 | 295.86 | 0.18% | |||
| 2007 | 314.07 | 317.53 | 1.10% | |||
| 2008 | 307.65 | 313.83 | -2.01% | |||
| 2009 | 303.75 | 304.79 | -0.34% | |||
| 2010 | ? | 323.86 | ? | |||
1 comment:
Thanks for sending that along Hank. All I can garner from this is that the Philanthromax seems to match up very closely with the Giving USA estimates (and I guess with the final Giving USA numbers for the years before 2008). So, what faith you choose to put in philanthromax really depends on how accurate you believe Giving USA to be. In order to really understand if the philanthromax has value we would need to know what the benchmark was that they were using when they built their algorithm – and I guess what I really mean is when they ran estimates back to 1969 what number were they comparing to in order to see if their algorithm was producing accurate output. If it was the GUSA numbers, then it makes perfect sense that it would be very close to GUSA number – now and going forward – since it was essentially reverse engineered to produce those numbers.
Post a Comment