Abuse of the Productivity and Innovation Credit: Problems and Mitigation

By Jeremy Chen
I recently got an e-mail from a certain “PIC Startup” promising to supply a “PIC Startup Package” comprising items like website development, search engine optimisation (SEO optimisation), social media marketing, a photocopier, this “PIC Startup Guide” which they charge $1,000 for independently, 10,000 flyers and a MacBook Air. The package costs $18,750 in total.
This is yet another on of an emerging niche of “consultants” that are trying to cream off a portion of tax payer funds disbursed in the Productivity and Innovation Credit (PIC). Let me talk a little more about this particular case. This “consultant” markets the package as something to help people who want to start a business to hit the ground running.

But yet on the landing page of their website (as of 9 Apr 2014), they write:
“Have you gotten your SGD 24,000 worth of PIC Benefits??”
And in bold red text “If you are earning SGD2,000 a month, this amount of cash grant is equivalent to 1 whole year’s pay!” Although all this is perfectly legal, this makes one deeply suspicious about the firm and the quality of start-ups that might emerge from such “packages”.
As such, PIC funds so disbursed are unlikely to create job-creating Singapore companies so much as give a small group a photocopier and a MacBook Air at tax payer expense. And since this “PIC Startup” firm presumably did their homework, it beggars belief that such basic items are regarded as “productivity/innovation expenditures”.
In this article, I would like to talk about the issues with the PIC scheme and suggest practical ways of improving its implementation.
The Productivity and Innovation Credit 
The Productivity and Innovation Credit was created to provide incentives for Singapore companies to invest in productivity improving technology. It is administered by IRAS and information on it may be found here and here.
What I feel to be the most positive aspect of the PIC is the 400% tax deductibility on qualifying expenses (as opposed to the usual 100%). The option to convert the expenses into a 60% cash payout was unsettling, but that may be argued to be a good way to help firms manage their cash-flow and risk can be perhaps managed.
For tax years 2013 to 2015, a shocking provision was added, the PIC Bonus. The IRAS webpage says that: “The PIC Bonus gives businesses a dollar-for-dollar matching cash bonus for YAs 2013 to 2015, subject to an overall cap of $15,000 for all 3 YAs combined.” (YA stands for “Year of Assessment”.)
 
The Problematic Cash Payout and the Inexplicable PIC Bonus
The PIC Bonus is what enables the “SGD 24,000 of PIC Benefits” mentioned above: Spend $15,000 on qualifying expenditure, get a 60% cash payout ($9,000) and a $15,000 dollar-for-dollar match. It does not make sense that there such a substantial surplus of 60%.
This issue was also put forth in a May 2013 letter to the ST Forum, where the writer expressed shock at the surplus, expecting the bonus to be limited to the 40% not covered by the cash payout of the PIC Scheme. Unfortunately, the response by IRAS failed to address that at all.
Cash payouts are double edged. The intention of the cash pay-outs are to help companies manage their cash flow, as cash flow might be a obstacle to making productivity improving expenditures because flexible payment schemes might not be available. (The pay-outs replicate a 400% tax deduction at a tax rate of 15%.)
However, what might arise is the inflation of invoices for cash payouts (the benefits being shared between the colluding parties). This is not a fanciful idea. It is a common form of fraud that arises because invoices can be minted by essentially anyone and the PIC cash payout option turns invoices into currency. Moreover, invoices are hard to audit for fraud. This means that catching fraudsters reliably would be difficult unless substantial effort is expended.
Moving away from fraud, there is the risk of companies making frivolous expenditures because it is cheap/free/more than pays for itself. Because claimants have little or no skin in the game, it then means that tax payer money is being poorly spent and will mostly go towards enriching the claimants and the vendors they buy from. With the 160% payout arising from the PIC bonus, the already cavernous wiggle-room for abuse (misuse and fraud) only gets bigger.
 
Addressing the Problems
I think the intent behind the PIC scheme (ex- PIC Bonus) is good. Forgoing some tax revenue each year is a reasonable trade for higher productivity in subsequent years. As such, I would like to propose a few measures whose application will reduce abuse. The measures do not depend on each other and are complementary. One has a very light footprint, while the other two are slightly more involved.
With this kind of policy design, the challenge is separation of the genuine cases from the fakes. In the history of economic thought, the idea that markets are imperfect due to information asymmetry is relatively young and has only just begun to sink in. (The 2007/2008 financial crash made many more aware of this.)
The central prescription of the theory has been that to separate genuine cases from fakes, one needs to observe a costly signal from the applicants, one that is relatively cheap for genuine cases and relatively expensive for others. The idea is that prospective fakes would be put off by the cost of generating the signal and while the genuine cases would be happy to do so.
In the first academic article on the topic, the example selected was the unknown quality of used cars, and the ameliorating mechanism that made the market work were warranties.
With the general principles in mind, let me talk about the possible solutions.
Simple Ratio Tests.
“Real” companies have real revenue and real operating expenses. Such companies also record these expenses in their accounts in detail. One low footprint screening procedure would involve obtaining revenue and operating expense numbers and considering the ratio of claims to revenue and the ratio of claims to operating expense.
There should be reasonable thresholds set for automatic acceptance. If one of both of the ratios is higher than the threshold, that should trigger some scrutiny. In cases, IRAS might choose to accept and process the claim, and in others IRAS might ask for more information. This is a cheap measure that should be implement.
Blacklists.
Productivity and innovation expenditures are expenditures that promote productivity and innovation. Insightful as that definition might be, it is clear that basic business expenses should not qualify. Examples might include: website development, photocopiers, advertising, and furniture. Computers, I guess, lie in a grey area. This has a very low footprint and should be implemented.
Cost Estimation.
The cost of a given type of item typically falls into a well-defined range. Knowing this range, requiring applicants to categorise expenditures and using this information to trigger scrutiny would be a strong measure against abuse. However, this requires cost estimation effort. On the other hand, that effort need only be made for categories that applicants frequently make claims in.
Alternative (Other than Cash Payouts) Means for Managing Cash Flow. 
The reason for offering cash payouts is to enable firms, which would otherwise find cash flow too tight, to upgrade their operations without “cash-constricting” those operations.
An alternative way of doing this would be for the government to enter into tie-ups with banks to facilitate short term loans at nominal interest rates (say 3% per annum or less), with the government paying the interest for up to 6 months if the expenditure is approved. In this way, companies that genuinely want to upgrade their operations will be able to do so and get the (larger) tax deduction.
Fraudsters, on the other hand, would find there to be too much skin in the game to try. Note a change of this sort has material effect on start-ups in their first three year of operations due to the generous tax exemptions available.) As such, this would “discourage start-ups from upgrading”. However, I would argue that start-ups should start up with new technology and methodologies rather than old economy ones.
The PIC scheme is good and well-intentioned, and has the potential to help “genuine upgraders”. However, the 60% cash payout and the dollar-for-dollar PIC bonus provide tremendous scope for abuse. What is more, detection of fraud can be extremely difficult and onerous. It would be far better to restructure the scheme such that “genuine upgraders” would not find it too expensive to navigate the checks while frauds would find it expensive to pass them so to ensure that public money is more well-spent.