By now, everybody, but everybody, has commented on the bad implementation of the demonetisation (DM) policy announced by PM Modi on November 8. In simple terms, R15.4 trillion (or lakh crore, hereafter referred to as T) were taken out of circulation on 8/11. Less than R3.5 T of new currency notes entered the system. Even by the end of December 2016, there will only be R6 T of new notes. Won’t this shortage of R7 trillion severely crimp transactions, diminish GDP growth and future tax revenues? These are important questions. Most reputable economists (both domestic and international) have opined that this radical demonetisation policy, while well-intentioned, was inappropriate and very costly.
In this discussion, we will take it as given that implementation of the policy has been terrible. But we differ with most on the expected short-run benefits, i.e., benefits without future policy changes. The latter we think are critical to the long run success of demonetisation. At the same time, we think the fear mongering about the short run costs needs to be discussed in a dispassionate manner.
First some broad “facts” to place the policy in perspective. GDP in 2016-17 was expected to be R150 T; R15.4 T of cash in circulation was withdrawn on November 8 (in the form of R500 and R1000 notes), as of November 8. As of November 27, R8.5 T had been deposited/exchanged with the banks, and unofficial rumours are that by November month end, almost R11 T of old notes were deposited.
Black cash (BC) in the Indian Economy
Currency with the public forms part of the transactions demand for money. While the demand for money question has been researched extensively by monetary economists, the demand for currency question has received only limited attention. The reason is obvious—typically, currency forms about 45% of the sum of cash and demand deposits (M1) in emerging markets, i.e., not big enough compared to demand deposits. In fiscal year 2014-15, cash in India was 62% of M1, a number that places it among the top 25% of emerging markets (EMs). (In 2000, the EM average was 36%, and India was 45%). This is our first crude estimate of BC in India—about 17% of M1 (difference between India and EMs: 62 minus 45), which, as of end October 2016, was R28 T, i.e., BC on November 8 was R4.8 T.
A more rigorous estimation of the demand for currency for purposes of consumption in EMs for the period 1980-2015 (with urbanisation, size of the agricultural economy, per-capita income etc. being part of the determinants) yields the following result—the demand for excess cash in India is the tenth highest among 60 EMs. Excess cash is defined to be the excess over that predicted by the currency demand model.
In the Indian sub-continent, the region most comparable to India in terms of culture (and black cash?), India is a major outlier, with Nepal a close second. On an average, black cash is a negligible percent of household consumption in the region excluding India and Nepal—actually, a negative 0.7%. The tenth most black cash economy in EMs is India (9.2%), and Nepal is eleventh at 7.8%. Interestingly, Thailand has more black cash than India, 10.3% of consumption; and Pakistan, considerably lower, at 2.7%.
Household consumption is estimated to be around R89 T in 2016, so black cash on November 8, in India (.092*89) was close to R8 T. We feel that this is a very robust estimate—and several policy implications follow this analysis.
Cost to GDP of demonetisation
The “belief” among “experts” is that black cash is only about 6% of the black economy. The black economy is variously estimated to be 25 % of GDP or around R40 T; the “popular” estimate of 6 % would place BC in India to be R2.4 T (what is the basis of the “6 % of cash is black” conclusion?)
In an article entitled “Monumental Mismanagement” (FE, December 4, 2016) former finance minister P Chidambaram vehemently argues that DM is a monumental failure. His reasoning: GDP will conservatively drop by 1% this fiscal year, and at least 90% of currency notes will return to the system, leaving the economy with zero-benefits from DM (GDP is R150 T, and notebandi cash is R15.4 T).
How accurate is the PC calculation likely to be? First, on GDP decline by at least 1%. Since all comparisons are with respect to 2015-16, what we do know is that agricultural GDP will be at least 5% higher this year. The Kharif crop was over by the time of de-monetisation, and the rabi acreage, despite dire forecasts, is proceeding at a 5% faster pace than last year. Industrial growth in November (PMI data) was also expanding in November. The one sector definitely affected by DM is the most important sector of the economy—services (60% of GDP). Like PC, assume conservatively that services consumption for the remaining five months of the year (November through March) of the fiscal year declines by 5%, with most of this decline coming in November-December. The service sector for five months (November through March) was expected to be R38 T. A decline of 5% is a loss of R2 T. Some of this (around R1 T) will be made up by 5% growth in agriculture, yielding a net loss of around R1 T.
So it does appear that the PC calculation is about right—i.e., the costs of DM are about equal to the benefits. But let us get back to our earlier calculations—R8 T is black cash, not black income stashed away in real estate, gold, foreign currency etc. Assume for a moment that black cash is the average of the two estimates—R2 T (“theirs”, the 6 % formula) and R8 T (ours, based on estimation of excess demand).
If R5 T is returned to banks with a tax penalty of 50 %, then that is a tax gain in 2016-17 of R2.5 T. At a lower 30 % tax rate in the future, this is an extra R1.5 T in perpetuity because these individuals will now be permanently in the tax net with a higher level of income than before. At a 5% discount rate, R1.5 T in perpetuity is R30 T. This is under the assumption that black cash has a low shelf life because smart money doesn’t keep black income in cash.
So a proper static benefit cost analysis of the DM policy is as follows. In 2016-17, costs are R1.5 T, and benefits are R1.5 + R2.5 or R4 T. Just a one year calculation shows a more than 200% return on investment! But future gains, through higher direct tax revenue collection of R1.5 T means a very, very large return on the investment. This result is predicated on the assumption that two-thirds of notebandi reflected legitimate income or savings, and only one-third was illegal. One can change various assumptions (amount of legitimate returns, discount rate etc) the result will not change. Even on a static no future policy change base, the de-monetisation policy will likely be a huge success.
But, if other policy measures to reform India are not forthcoming (decline in personal income tax rates to discourage tax evasion, elimination of stamp duty and long term tax rates on property, a clampdown on the extortion power of tax officials, and reform of election funding) then history will view demonetisation as a colossal failure of thought and reform. It would be so unlike PM Modi to stop here, and I am not betting that he will.
The author is contributing editor, The Financial Express, and senior India analyst at Observatory Group, a New York-based macro policy advisory group. Twitter: @surjitbhalla, Views are personal.