In 2014, Alexandra Humphrey Cifuentes, Jonathan Pryke and I published the Devpolicy Discussion Paper, “Putting our money where our mouths are? Donations to NGOs and support for ODA in Australia”. The paper compared 2012 data from the ACFID Electorate Snapshot Survey of donations to aid NGOs, 2013 Vote Compass data on public opinions about ODA, election results and 2011 census data. Earlier this year I was advised of probable issues with source data that we had drawn on for the paper.
The statistical analysis we had conducted in our original paper was technically correct, and we had used the source data appropriately. However, there were problems with the source data. And so, because we were no longer confident we could trust the data, we were no longer confident in the paper’s findings.
Accordingly, I started data-gathering anew, combining new 2015 ACFID Electorate Snapshot Survey data on aid NGO donations, 2016 Vote Compass data on public opinion about aid, and 2011 Census Data. As I did this, I was able to get much better data thanks to the active assistance of ACFID as well as the Campaign for Australian Aid (the Campaign partnered with us to purchase the Vote Compass data in an improved format). With a new dataset, I re-analysed the data, and Alexandra, Jono and I have just published a new version of our original paper.
Because of the data concerns with the original paper, we’ve chosen to retract the 2014 paper and replace it with our 2016 analysis. The new paper can be downloaded here. (All Devpolicy papers are available from this page.)
Many of our original 2014 findings remain unchanged. Green party support and academic education are still positively associated with the proportion of electorates’ voting-aged populations who want to see ODA increased, and they are also still positively associated with the proportion of electorates’ voting-aged populations who gave to aid NGOs. Most significantly, we still found a positive relationship between support for ODA increases and donations to aid NGOs. Australian electorates where support for increased ODA was highest also tended to be electorates where a higher share of the voting-aged population donated to NGOs. You can see this in the scatter plot below (each point represents an electorate). (For all the technical details please see the paper itself, or email me.)
Support for ODA versus share of population donating to aid NGOs
There were also some changes though. Some were minor or not particularly interesting. However, one change was of interest. As in the original paper, there was a positive association between support for Labor and support for ODA increases. Similarly, there was still a negative association between support for the Coalition and support for ODA increases. (All of these findings come from analysis in which sociodemographic variables are controlled for.)
However, in the new version of our research we found a positive association between support for the Coalition and the share of electorates’ populations donating to NGOs. The substantive magnitude of this effect was small, but it was statistically significant. We also found a commensurate negative association between support for Labor and the share of electorates’ populations donating to NGOs. You can see this in the figure below. The top two charts are for the Coalition. The x-axis on both charts reflects the proportion of an electorate who voted for the Coalition. The y-axis of the top-left chart reflects the share of an electorate’s population who we would expected to donate to NGOs on the basis of our regression results. The y-axis of the top-right chart reflects the share of an electorate’s population who we would expect to support an increase in ODA.
The bottom two charts are identical except that they plot results predicted on the basis of support for Labor.
Support for Labor and the Coalition and its association with support for ODA and donations to NGOs
As you can see, the donations line slopes upwards for the Coalition and downwards for Labor. On the other hand, the ODA line slopes upwards for Labor but downwards for the Coalition. Also obvious is that the party effects on ODA are much larger than they are on donations. Electorates with more Labor voters may a have a few less NGO donors, but they have many more ODA supporters. Electorates with more Coalition voters may a have a few more NGO donors, but this is more than offset by a decrease in ODA supporters.
For (much) more detail have a look at the paper. Thanks to the help of our civil society partners, we think the 2016 version of the paper is a big improvement on the 2014 version. There’s still more we’d like to do to further advance research in this area. This will require additional years of data though. For now, the findings you’ll read in the new paper reflect the state of our knowledge.
Terence Wood is a Research Fellow at the Development Policy Centre. Terence’s research interests include aid policy, the politics of aid, and governance in developing countries.