Transparency in Municipal Budgets - or does PAIA work?

What percentage of PAIA requests get answered by municipalities, and is this related to any specific municipality characteristic like size or population composition. Secondly, is how you ask for data important?

Transparency in Municipal Budgets - or does PAIA work?

I don’t know too much about open data, although if your husband is Adi Eyal, you tend to pick up a few things. I’m an Economist at the School of Economics at UCT – which means that while I’m very good at working with data, and being able to evaluate fairly well whether an impact can be considered a correlation, or a causal effect – I don’t know that much about transparency, accountability etc.

All that said, thanks to a very dedicated and smart student of mine, Stephanie van der May, I do now know a little bit more about the efficacy of the PAIA (Promotion of Access to Information Act) process, particularly when it comes to municipalities.

I’ll try to keep the description short. Stephanie and I decided we were interested in two things. The first was descriptive – what percentage of PAIA requests get answered by municipalities, and is this related to any specific municipality characteristic like size or population composition. The second was analytical. We wanted to know whether how you ask for data is important. Does tone matter? And can we pinpoint the most effective manner of approaching government bodies to obtain the data we want?

We devised a very simple experiment. Stephanie sent PAIA requests to 223 municipalities, divided up into requests we deemed aggressive, and requests deemed neutral. Allocation to either group was random, and not correlated with any municipality characteristics. For the aggressive requests, we asked Gabriella Razzano from the Open Democracy Advice Centre for appropriate wording. Using the language from the PAIA Act, we asked for our data in the most aggressive way possible. We presumed that we would get a refusal, and in the original request we reminded the municipality of the penalties for refusal, and the possible consequences. The other requests were deliberately neutral in tone.

The data we requested should have been easily available to municipalities, and the process of fulfilling our request should have taken no longer than ten minutes. We asked for the municipal annual budget, and certification that the budget has been locked (finalised, and thus must be adhered to). It is mandatory for the budget to be uploaded to the municipality website (although our survey showed only 50% of municipalities have actually done so), and the budget locking certificate must be submitted to Treasury by law.

What did we find? We obtained 32 responses. A little over ten percent. We sent an appeals round – we got 13 more responses. The aggressive appeals were even crosser this time around – you’d think someone would be scared by being threatened to be taken to Human Rights Commission.

To our other question – does aggression work? Well, with such a small sample size, and so little variation in the response variable, we would be irresponsible to claim any strong conclusions here. But we can tell you that aggression has no impact on whether municipalities respond, but if they do respond, aggression shortens the average response time.

So were our responses vanishing into thin air? We used the most up to date list of information officer details for municipalities (March 2014). To cover our bases, we emailed the information officer, mayors, and other municipal officers. We made an awful lot of phone calls to verify addresses, and we checked why requests weren’t being answered for a small sub-sample. These checks revealed that many municipalities didn’t know who their information officer was. Many requests are handled by the personal assistants of municipal managers – and clearly PAIA doesn’t feature high on their list of priorities.

I’ll let you decide what the main conclusions are here, but I hope that we have added value to the debate. If you’d like some data from a municipality, I hope you have a contact there, and I don’t suggest going the PAIA route – you won’t find much joy. Don’t bother getting angry – it doesn’t make a difference.

The paper with full details of our methodology etc. is available on request – mail Katherine.eyal@uct.ac.za (You don’t need to fill out a PAIA request). To mail Stephanie and commend her on her work, mail stephanievdmey@hotmail.com.

Share this post:
Email iconTwitter icon

What percentage of PAIA requests get answered by municipalities, and is this related to any specific municipality characteristic like size or population composition. Secondly, is how you ask for data important?

Transparency in Municipal Budgets - or does PAIA work?

I don’t know too much about open data, although if your husband is Adi Eyal, you tend to pick up a few things. I’m an Economist at the School of Economics at UCT – which means that while I’m very good at working with data, and being able to evaluate fairly well whether an impact can be considered a correlation, or a causal effect – I don’t know that much about transparency, accountability etc.

All that said, thanks to a very dedicated and smart student of mine, Stephanie van der May, I do now know a little bit more about the efficacy of the PAIA (Promotion of Access to Information Act) process, particularly when it comes to municipalities.

I’ll try to keep the description short. Stephanie and I decided we were interested in two things. The first was descriptive – what percentage of PAIA requests get answered by municipalities, and is this related to any specific municipality characteristic like size or population composition. The second was analytical. We wanted to know whether how you ask for data is important. Does tone matter? And can we pinpoint the most effective manner of approaching government bodies to obtain the data we want?

We devised a very simple experiment. Stephanie sent PAIA requests to 223 municipalities, divided up into requests we deemed aggressive, and requests deemed neutral. Allocation to either group was random, and not correlated with any municipality characteristics. For the aggressive requests, we asked Gabriella Razzano from the Open Democracy Advice Centre for appropriate wording. Using the language from the PAIA Act, we asked for our data in the most aggressive way possible. We presumed that we would get a refusal, and in the original request we reminded the municipality of the penalties for refusal, and the possible consequences. The other requests were deliberately neutral in tone.

The data we requested should have been easily available to municipalities, and the process of fulfilling our request should have taken no longer than ten minutes. We asked for the municipal annual budget, and certification that the budget has been locked (finalised, and thus must be adhered to). It is mandatory for the budget to be uploaded to the municipality website (although our survey showed only 50% of municipalities have actually done so), and the budget locking certificate must be submitted to Treasury by law.

What did we find? We obtained 32 responses. A little over ten percent. We sent an appeals round – we got 13 more responses. The aggressive appeals were even crosser this time around – you’d think someone would be scared by being threatened to be taken to Human Rights Commission.

To our other question – does aggression work? Well, with such a small sample size, and so little variation in the response variable, we would be irresponsible to claim any strong conclusions here. But we can tell you that aggression has no impact on whether municipalities respond, but if they do respond, aggression shortens the average response time.

So were our responses vanishing into thin air? We used the most up to date list of information officer details for municipalities (March 2014). To cover our bases, we emailed the information officer, mayors, and other municipal officers. We made an awful lot of phone calls to verify addresses, and we checked why requests weren’t being answered for a small sub-sample. These checks revealed that many municipalities didn’t know who their information officer was. Many requests are handled by the personal assistants of municipal managers – and clearly PAIA doesn’t feature high on their list of priorities.

I’ll let you decide what the main conclusions are here, but I hope that we have added value to the debate. If you’d like some data from a municipality, I hope you have a contact there, and I don’t suggest going the PAIA route – you won’t find much joy. Don’t bother getting angry – it doesn’t make a difference.

The paper with full details of our methodology etc. is available on request – mail Katherine.eyal@uct.ac.za (You don’t need to fill out a PAIA request). To mail Stephanie and commend her on her work, mail stephanievdmey@hotmail.com.