Giving money away is not easy, even if it is to the government in lieu of essential services
Neelkanth Mishra | Last Updated at November 6, 2017 23:28 IST
(This was published in the Business Standard: link)
Crowd control is as much about psychology as it is about hard policing. If 5 per cent of the drivers on a city’s roads do not follow traffic rules, they can be caught and penalised. However, if half or more of drivers do so, there is very little that traffic police can do. Very quickly the remaining half stop following rules too as rule-following is costly for the individual: Not only is there no resultant benefit if everyone does not comply, the probability of getting caught falls as more people break rules. This is one of the reasons why individuals behave differently in mobs than when on their own.
Tax payment is no different. Giving money away is not easy, even if it is to the government in lieu of essential services. Citizens and enterprises balance their sense of duty and nationalism against the probability of getting caught for non-compliance (no way can a few thousand tax officers police tens of millions of tax payers), and perhaps more importantly, of fairness (“Why is not everyone paying taxes; why should I be the only one doing national service?”).
At the start of the goods and services tax (GST), the anecdotes were encouraging. An improvement in upstream compliance was anyway expected: That is, companies would make their suppliers comply; they in turn would do the same to their suppliers, and so on. But a positive surprise was that many firms were also enforcing downstream compliance, which is, getting their customers to register: Wholesalers were refusing to supply to retailers who did not have a GST number. This may have been because of the belief that the GST was positioned and branded as a fool-proof system, and that the government would be able to monitor businesses at a level of detail never seen before. An equal if not more important driver may have been that everyone was doing it. Further, demonetisation had shaken up the sense of impunity with which people evaded taxes.
But the complexity of filing GST returns appears to have stalled this progress, and now threatens to reverse it.
Inherent in the design of the GST return-filing process is an assumption that most GST returns would be automatically generated by software into which businesses anyway enter transactional data. There is nothing wrong with this in principle: I learned early while working as a data warehousing architect that to ensure good data quality, one must remove manual intervention. However, in the case of the GST, it appears that only 1 to 2 per cent of the enterprises have both online systems integrated with their businesses, as well as tax consultants who can prepare the returns mandated by the GST. The rest use tax accountants that are given summarised data, generally have offline systems (that is, data is entered much after the business transaction is over), and many are used to maintaining multiple books of accounts (in an attempt to evade tax). Hard as it is to avoid manual intervention in such situations, the complexity of some of the tables that must be filled up has steepened the learning curve enterprises must go through.
The challenge becomes clear from the fact that firms struggled to upload their sales details for July despite being given 70 days to comply. This same period for August is coming to a close and there is no deadline being set yet, implying it would take even longer. To put things in perspective, when the GST stabilises, these need to be completed within 10 days of the month-end. Matching of purchase details from companies with sales details of their suppliers is the core feature of the GST, the fear of which is expected to improve compliance. However, that process even for July is far from complete, with only a few hundred thousand firms submitting purchase data.
These are not just data issues: Many firms have refused to pay for purchases till the supplier has uploaded the invoice in the system, and thus feel reassured about the input-tax credit. The more the delay in filing, the longer the payments remain stuck, and working capital demands increase. Once matching starts, and input tax credit is denied, or in some cases where suppliers may have uploaded more than what their customers wanted to disclose, there is likely to be friction between supply-chain partners. Worse, as the months roll by and the backlog piles up, many firms might just refuse to file any returns.
Should businesses go down that path, getting the crowds to behave again will require significant effort, and the system would be thrown back many months if not a year in terms of broadening the tax base and improving compliance. Even if temporarily, simplifying the returns (for example, are the HSN codes and their unified quantity code really necessary right now for all items being sold?) and their upload would make filers fear the GST less. This may also need changes to GST rules: Limiting all changes to just software may not suffice.
The effort should be to nudge smaller enterprises into entering transactions online into systems as they happen, and then filing returns seamlessly. As the costs of maintaining multiple accounts and paying a tax accountant become obvious, such a transition should happen naturally. This would, however, take a few years: Changing the behaviour of tens of millions of businesses cannot be quick. For many, the GST first seemed to be about just buying a computer: That they need to invest in the appropriate software and then integrate it with their business is only just becoming clear.
As the GST Council debates new rules, a road map for the transition would be helpful: While the final vision is encoded into law, a path with milestones is missing. The invoice-matching feature though should be retained (some have suggested that it be replaced with statistical sampling of returns). That is the stick that together with the use of analytical software can help tax-administrators project more power than they have. This is important, just as the use of cameras at traffic signals in Mumbai, combined with automated challans messaged to errant drivers has improved driving behaviour. The fear that someone may be watching the camera feed has led to people stopping at red lights even if there is no traffic policeman in sight.