The Government of India (GoI) has an informative website with all budget speeches and the complete set of budget proposals. Consider these simple questions:
1. Which department got the highest outlay?
2. What has happened to expenditure on child-related schemes over the last 10 years?
3. Which states are getting money for laying roads?
4. Which states have not used allotted money for primary education?
5. Do reserved constituencies get less allocation?
These are reasonable questions people would want to ask of the budget. The information needed to answer the first two is there, but it would require us to open dozens of PDF files and compute digitised information manually! The first question is rather simple, but to answer this I have to manually add plan and non-plan outlay for each department and then find out which department got the largest allocations. This is because all information is presented in PDF files that are good only for printing.
Providing it in a ‘machine readable format’ would enable us to answer many questions with ease that we cannot do with PDF.
Bane of aggregate information
Budgets are made by adding up demands from each state, but only totals are available for our use. Thanks to this, a simple question like how much money has each state got for laying roads or to start child-care centres becomes unanswerable. If the entire budget process is digitised from source systematically, we can get primary data with which answering such questions will become possible. It will also make it easier for the officials who work on this process.
A Dalit politician told me recently that reserved constituencies get lower budgetary allocations. This is a serious question that should examined, but practically cannot be done with budget information. When the finance minister tells us how much money was used last year by each department, it is based on information that was recorded for every rupee spent by each department in each village. Today every village does its accounts and these are replicated by the district, and by states and then the centre. Such duplication can be avoided by digitising at source and developing a system for sharing that information upwards.
Suppose we can get village-wise expenditure, even then we cannot find if reserved constituencies spent less money. If we have another database that says which villages belong to reserved constituencies, this task will become simple. Similarly having geographic databases can help us see this information on a map to see regional variations. No database including a detailed budget can help us answer all important questions. The ability to combine information from other databases can be critical to answer important questions, and such analysis can be insightful and useful in democratic engagement. In order to do so, there should be a commitment to digitise census, maps, and other basic information and make them available freely for public use.
One of the 8 principles deals with information being accessible for the widest range of purposes. Suppose there is an active people’s organisation in a district that wishes to track all policy issues relating to the district. By designing all government data in accessible formats such as XML, the group can easily combine information from many sources in a convenient format. For example, it can create a report that combines what questions their MLAs and MPs asked in assembly and parliament, news reports about the district, how many people got employment in NREGA in the last quarter, which funds were not used by the administration, etc. Given the explosion of online information, a lot of information to answer these questions is available online. But thanks to poor formatting of information, such reports will have to be done manually.
Government of India’s budget website is the good example of an important and informative website that is nearly useless thanks to the poor formatting of information. Making the same information available in machine readable formats will take us a long way. The real revolution lies in creating systems to make budget to the last detail available to any user.