INTRODUCTION
The Estimates committee (1980-81) in 7th Lok Sabha, Ministry of Finance had recommended that government should attempt an estimate of black economy to formulate appropriate policies. The study of unaccounted income thereafter was entrusted to the National Institute Of Public Finance And Policy in July 1982 to: identify sectors generating black money, causes of such, studying the methods employed to generate black money and channels through which concealed income is invested and spent in other ways, methods employed to convert black money into white, broad estimate of money generated and to take regional or sectoral surveys required in connection with the above. The Institute submitted its report in March 1985.
The focus of the paper is on methods and measurements of estimating black economy and therefore, other aspects of the report like distinguishing between black income and black wealth, giving an explanation for the causes of black economy and remedial measures etc. are not explicitly dealt with, yet these issues are kept in mind and the methods and measurements are discussed at the level best. This paper has tried to do some justification to the NIPFP REPORT (1985) and the comments by various known scholars.
METHODS FOR ESTIMATING BLACK INCOME
The first approach is the fiscal approach: Variants of this approach attempt to arrive at independent estimates of incomes subject to tax and compare these with the income actually assessed for taxation and call the difference between the two- a measure of tax evaded income. N.Kaldor first used this method in 1956. This method was also used by Wanchoo Committee report and also used by Chopra (1982) to estimate series of unaccounted income in India. Such studies have also been conducted in United Kingdom by Kenadian (1982) and Park (1981). This method is discussed in detail below.
The second approach is the monetary approach which assumes stability in relationship among various money stock aggregates among each other and to the total transactions in the economy. On a broader note , there are three categories of monetary approach. Gutmann in 1977, first pioneered an approach where he picks a base year which has negligible size of unaccounted economy. For this particular year it takes the C/D (currency to deposit ratio) fixed and any subsequent increase in this ratio is attributed to disproportionate growth in demand for cash finance in unaccounted economy. Yet, this case is non-sensical in the sense because C/D ratio has fallen in India accounting for ‘negative Black economy’. Second variant is applied by Gupta and Gupta in 1982 using the same methodology of ratio of total monetized transactions to total nominal GNP. Third method developed originally by Cagan and later developed by Tanzi specifies a currency demand equation with an included tax variable in the explanatory variable, used in North America and Europe.
The third approach is Physical Input approach which makes use of the Input – output matrix. An intermediate input whose consumption can be measured from reliable data and the same input can be explained in terms of value added growth in officially measured GNP. The ‘residual consumption’ is measure of unaccounted economy.
The labour market approach, which is precisely the fourth method, can be gauged from official labour market participation rate, if these are inexplicably low compared to the countries where the black economy has limited significance. Example,in Italy. Relevance of this method in India is limited because of difficulties in employment data (a person in India can be self employed and at the same time can be engaged in a profession that can generate wage income).
The fifth approach is National Accounts approach relies that a country measures GNP all non- from both Income and Expenditure side. Such independence of income and expenditure is not complete in India (because the private final consumption expenditure rely on estimates of gross output by sector of origin and the latter are linked with the estimate of value added by sector of origin).
The report of NIPFP (1985) estimated the black economy applying the fiscal approach, so a great deal of attention has been given to this approach. This is discussed below.
Taxonomy of alternative methods for measuring the black income, is given below-

FISCAL APPROACH
An overview of the key assumptions and steps in this method are as follows-
The starting point is- estimating Income from sector of origin from national income accounts. Assuming the fact that there is no question of tax- evasion in agriculture and in all other sectors assuming salary incomes are fully reported for taxation, for all non- agricultural sectors, the ratio of non- salary income to total income is estimated. The proportion and amount of non salary income above the income tax exemption limit is estimated and summed across all the sectors which yields non- salary income assessable to tax. This is subtracted from the actual non-salary income assessed for income taxation. The difference gives the estimate of tax- evaded income.
National accounts statistics which is the starting point is used to derive the statistics for gross –personal income accruing to households. CSO provides data on GDP, GNP and NDP: equally important is the fact that how this income is DISTRIBUTED across the sectors because had the gross personal income been distributed equally across all earners then none would have been under the taxable income tax in the year 1975-76 and 1980-81. The distribution of gross- personal income is provided by NCAER- National council of applied economic research, starts with a primary survey and employs a multistage sampling strategy. This survey was skewed to achieve over-representation of low and middle income households for the year 1975-76 and separate tabulation of rural and urban households was desirable since bulk of income taxation comes from the rural households. Assuming there is one to one correspondence between amount of income and the number of households who earned it, still the NCAER data suffers from the drawback that it includes households and not earners. For exploring the taxable assesses it the earners who are important. Criteria for converting household into earner are complex. The single- earner household case is easy. For multi- earner household , the concept of ‘ average earner’ is taken as the norm and assuming the average income per earner is also the actual income for all earners in the relevant income range. A derived frequency distribution for earners is then obtained for rural and urban income earners. Clearly, the average norm violates reality and is a possible source of error. Further, we note that frequency distribution resemble typical distribution for wealth, income, consumption in which large number of earners are concentrated at relatively low levels of income.
The NCAER based survey falls substantially below the NAS estimate. How do we account for this discrepancy? First, note that planning, investment allocation, regional policy, fiscal and monetary policy, all rely on NAS. The household’s survey yields lower totals for macroeconomic aggregate than national accounts especially for developing country. This divergence is attributed to under-reporting in the fear of fiscal consequences. The missing income is then decomposed into urban and rural earners, the degree of under-reporting is assumed to be uniform across all income range. Further, since the opportunity of reaping the black income is higher among urban earners, in the similar manner, under-reporting is high because of the progressive tax structure. Yet, the fiscal approach takes into consideration how to allocate the missing income into different degrees of under-reporting among urban and rural areas and take the middle way rather the extreme way of allocating the entire missing income to the urban area.
Given the derived frequency distribution (which is converted to lognormal), allowances are made for exemptions, exclusions and deductions under Income tax- act. Noting that, once household income has been converted to household income ranges, in the same way earners have to be converted into earner income range. This step is taken because exemptions and deductions pertain not to income in general, but to source of component of income in general. A simple case in point, says agricultural income are exempt from tax, but full exemption does not extend to livestock income. In the same manner, HRA paid to salary earners by their employers is exempt from tax subject to certain specified limits. The provident fund and business depreciation included in the gross personal income is also subject to exemption under the Income Tax Act. Correspondingly, these deductions are made from urban and rural India. Thus, what we have is disaggregated source components, regrouped into earner wise income ranges which are consistent with frequency distribution by earner wise income ranges, this is the gross personal income of earners assessable to tax. On this the further, exemption limit is allowed for.
The next step calls for estimating the tax evaded income. Since the report of NIPFP (1985) takes two assessment year into consideration, (i.e. 1976/77 and 1980/81) approximations have been done in the publications of AIITS and C.&A.G. report which provide information on income assessed to tax on assessment year basis and number of assesses respectively, adding or subtracting capital gains, refunds etc. In the given report , for the assessment year 1975-76, the tax evaded income as a percentage of GDP( at current prices) for the non- corporate sector comes out as 3.7%( after adjusting for refunds) and 5.6% (after adjusting for capital gains). An important assumption made between the years 1975-81 is concentration of incomes remains unchanged.
Considering that the above estimates are made on the assumption that the National Income estimates are accurate, it could be well possible that aggregates themselves are distorted by evasion behavior. Say when under-voicing of production or over-voicing of costs takes place in manufacturing sector, then it leads to under-estimation of value- added in the sector. Thus there is a strong prima facie case for suspecting significant under-estimation of total GDP.
Critic (support) of (for) fiscal approach
The NIPFP report estimated the size of the black economy to be 20% of white economy for 1980-81 using this method. S.B. Gupta pointed out the errors and pointed out the black economy to be 42% of GDP. He critically evaluated the NIPFP estimate and said it contains “ an unknown degree of bias”. But, without establishing and reassessing the bias, he uses it in his estimates along with the addition of indirect taxes like excise duties and corporate tax which leads to multiple counting of black incomes. It is not doubted that size of public sector in the National Economy is significant, so any leakage which takes place from this sector should be taken into consideration, which is not considered in both the above mentioned estimates. Sonali Basu in 1996 corrected the bias and estimated the size of the black economy to be 25% of GDP for the year 1980-81, if capital gains are to be excluded (It is commented below that capital gains or any kind of transfers are not part of national income).
Now, it is worthwhile to note that all the above mentioned estimates is based on the definition which is itself questionable- ‘income which are taxable but not reported to the tax authority’. This definition does not discriminate between legal or illegal income, it implies illegal income that are below taxable limit are left out. Yet this method calls for institutional details, that is studying components of individual sector that involves specific requirements.
MONETARY APPROACH
As mentioned in the introduction, there are three variants of monetary approach. The NIPFP report discusses the method adopted by Tanzi in Indian context. Basic element of Tanzi’s method is; major transactions in an economy are carried out with currency. So, size and growth of unaccounted economy directly influences public demand for cash. It isolates that part of public‘s currency which is illegal and then on basis of its assumption about income velocity of such currency the size of the unaccounted economy is identified. Following equation is formed- ln (C/M2) = a0+ a1ln T+a2ln (WS/NI) + a3lnR where C/M2 represents ratio of currency to broad money, T is the tax variable , WS/NI is the share of wages and salaries in National Income, R is the rate of interest on time deposits and Y is the real per capita GNP. From this equation, currency demand is obtained (Ct), given the observed value of the other variables in that particular year. Another value of currency demand is obtained by setting the value of tax variable equal to zero. The difference gives us the holding of money due to say increase taxation. This is then subtracted from M1 which gives the ‘legal money’. Now noting, Nominal GNP/legal money gives us the income velocity for legal money. Tanzi assumes that this is also the velocity for illegal money for the year t.
A lot of changes were formulated when Tanzi’s method was adapted to Indian system. Instead of C/M2 , C/P was used to focus on public’s demand for real cash, T was replaced by T/Y which includes all kind of tax revenues accruing to centre, state and union territories since in India 80% of tax revenue accrues from Indirect Taxes and thus evasion of commodity taxes could generate powerful cash demand. The effect of changes in the expectation of price level on currency demand is also taken into account. The equation now becomes-
C/P= a0+a1RGNP+a2NOC+a3TTR+a4(R-ERI) +a5ERI
Where RGNP is real GNP at 1970-71 market prices, NOC is number of commercial bank branches, TTR is ratio of total tax revenue to current market price GNP, R is nominal rate of interest on bank deposits and ERI is expected rate of inflation. Assessing the signs of coefficients, a1 has a positive value because the demand for real cash balances increases as country’s real income increases: a2 will have a negative sign because the more the number of bank branches, it is easier for people to hold less cash and shift portfolio towards bank deposits: a3 will have a positive sign, as the burden of tax increases, people will be tempted to evade more tax: a4 will have a negative sign, as the real rate of interest increases economic agents will substitute interest bearing deposits for currency: a5 will also have a negative sign, because as the expected rate of inflation rises, people will switch to real goods.
Critic(support) of (for) Monetary approach
When the equation was analyzed when adapted to the Indian system, it was found to have good statistical properties like adjusted r2, Durbin-Watson and F- statistic, except the sign of a2 was found out to be opposite because of studies of NOC absent from RBI data. This method was also adopted by Gupta & Gupta.
However, many argue that there is lot of inherent flaws in Tanzi’s approach. First , there is an implicit assumption that none of the ‘illegal currency’ is used to finance transactions relating to recorded GNP and estimated stock of legal money is related to measured GNP. But then, unaccounted income should be wholly added to measure GNP or some part of it may be captured. Therefore how can the legal money which measures GNP and which in itself is determined by income velocity of ‘legal money’ be logically different from illegal money? Second, the income velocity for illegal and legal currencies cannot be the same as argued by many authors. Some may say that the velocity is lower in underground economy than in the legal economy while some argue for the opposite. Often , the income generated is a small part of total transaction like in purchase of Gold or real estate, in these cases transfers account for a major portion of value of transaction. Since volume of such transactions has increased, so has the velocity of circulation in these transactions.
Third, income velocity of illegal currency might not be necessarily stable, say when indirect tax increases, it leads to additional demand for currencies in order to conduct transactions, it is quite possible that increase in income could occur without evading tax.
Fourth, Feige argues that tax evasion may also occur through banking system instead of only cash. Fifth, this method is ill suited for under-reporting through non tax ventures.
EDGE OF FISCAL APPROACH OVER MONETARY APPROACH
Through this entire discussion over the two approaches mentioned above, there are many clarifications which can be sought through. One of them being, the fiscal approach provides room for institutional details because it recognizes that different method should be adopted to study each sector (Kumar, 1999). We saw that fiscal approach disaggregates the household into components of sectors, which are then further classified into income earner range; on basis of this the tax evasion is assessed after accounting for exemptions and deductions. Contrary, the monetary approach just takes into account those sectors which uses money as transaction and misses out those which does not. Now we saw, fiscal approach provides the leverage to also gauge out the income which are below the taxable limit and then consider the rest of the income to be assessed for tax, but the monetary policy miss out such ‘ below tax limit’ and treat it as black income. Therefore, one would expect the estimate of black income to be higher under the monetary approach.
PHYSICAL- INPUT APPROACH
The first step towards this measure is to take an intermediate input for which reliable data could be obtained, say electricity. The approach is to identify a stable relation between use of electric power and National Output and then see if the growth of officially measured GDP can account for growth in electricity consumption. The main steps are as follows: Firstly, this method assumes a fixed linear relationship between the intermediary input, electricity in this case and total value added ,consisting of both reported and un reported. Thus , for a year ‘t’ say. λ = INt/TYt
Defining a variable βt which equals the ratio- total GDP/ Reported GDP
Therefore, INt = λ .βt. Reported GDP ( for year t)
Writing λ .βt = ℇt
Changes in technology and output mix are also accounted for. Say, ITt and IPt represent proxy variables respectively, then resulting equation is-
INt = ɑ + ℇtRep GDPt +γ1 ITt +γ2 IPt
Noting that, the ℇt can change overtime because of changes in βt , thus alternative method for estimating ℇt is
ℇt = ℇ0+ℇ1t+ℇ2t2+……….
Gupta and Mehta generated estimates for the unreported economy as follows- YEAR | % of total GDP | % of reported GDP | 1964-65 | 2.7 | 2.8 | 1974-75 | 12.1 | 13.8 | 1978-79 | 16.4 | 19.8 |
Critic (support) of (for) physical input approach
As clearly seen from above, ℇt is a product of two parameters but it is only βt which gives an estimate for unreported economy. This implies one requires independent knowledge of the value of λ or unreported economy is non- existent in some base year, which is not clearly defined by Gupta & Mehta. It is also to be noted that λ is also defined poorly for variety of reasons. Say for example, different energy sources have different efficiency level: energy as an input may differ and can be substituted from one form to another like bio organic materials to kerosene , coal etc.. to solar power.
There are further arguments against this estimation like that of assuming a fixed coefficient relationship between power consumption and National output. Some light is shredded on this issue. Noting that this linear relationship might hold true in the industrial sector, but say we are applying the same method in the service sector, wherein trade can expand with relatively little change in demand for electricity. Moreover, at higher level of aggregation , the variable of technical change is incapable of reflecting the output shifts within the broad sectors as statistics turns out insignificant (i.e IPt incapable of explaining whether output changes due to increasing rural electrification or increasing final consumption).
A second flaw in this method is final demand for electricity is affected by variables like pricing of electric consumption, change in income etc. which can affect the value of λ, and this would undermine the relation between input- output mix. Further, it is not necessary that electricity production is always equal to electricity consumption, because of transmission losses or theft so what is sold and what is actually used are two different amounts.
NATIONAL ACCOUNTS APPROACH
This approach was undertaken by Ghosh, Rastogi, Bagchi and Chaturvedi in 1981. The paper analyzed trends in capital formation, growth of domestic product and capital- output ratio. This paper argues that a high rate of capital formation is not reflected in higher output growth. For explaining such a phenomenon, the authors dwell on the possibility of underestimating official data on GDP, their focus is on the investment – output puzzle. They do not deploy any complicated methodology but rather examine the accounts by sector and suggests some magnitudes by which output and value added might be under-recorded in some key sectors.
They hazard the gross value of output from manufacturing sector has been understated by 10% for tax evasion reasons. Similarly, the gross value added in trade and other services has been understated by 15%, municipal valuations are looked upon for rental and housing which definitely underestimates rent by 20% because of underlying rent- control laws. Combining these assumptions they estimated that unreported GDP must have been 7-9% of current market price in the years 1970/71 to 1977/78.

Critic ( support) of (for) National accounts approach
This method is based on the simple fact that there must be an earning to every unit of expenditure. Rather than cropping up a single estimate for the underestimated GDP which is frequently seen in the newspapers and magazines, this method provides three different guess estimates for three sectors. Yet, this method is the informal of all the methods mentioned so far. Ultimately, these estimates are guesses and they are not supported by any quantitative information, it reflects informal judgment.
NSSO undertakes consumption surveys also by classifying the population according to consumption classes, but such estimates are likely to reliable. The lower income categories resort to consumption loans and consume more than their income. The upper income categories are inadequately covered in the survey often reporting biased consumption. As one moves up the income ladder, it represents their income less & less (Kumar; The black economy in India).
LABOUR MARKET APPROACH
This approach has mainly been used in Italy. Official labor market participation rate can account for under-ground economy, but as mentioned earlier relevance of this approach in India is limited because of difficulties in employment data. A person might be self employed and a salaried official at the same time. In rural areas, migration is common; people are casual & seasonal employees or daily wage earner or sometimes a combination. Thus it is very difficult to estimate data on labor market participation rate with often children being part of it especially in the construction sector. Even though one includes a guess estimate of women and men in the labor force, yet there is discrimination of wages in the informal sector against children and women. Inevitably, over invoicing of cost and under invoicing of revenue will be consequence of such reports.
SALIENT FEATURES OF THE STUDY
→Studying the methods for estimating black economy has been along following lines: To formulate an appropriate conceptual design for estimating the black economy, to explicate the methods used for converting the black income into white income and thus scaling down black economy. Why scaling this down is important? A close approximation of the amount of black income is important from policy perspective, to help achieve socio- economic goals.
Comment by K.N. KABRA
But, as a critic of the report, professor K.N.KABRA says that this report has not distinguished between policy instruments like taxation, regulation and economic controls and policy objectives, which was supposed to be the aim of this report. This then implies that problem of black economy is not placed in context to overall growth strategy.
→To make situations worse, the caveat is placed by some authors that change in personal incomes have repercussions on the true and recorded national income while some of these incomes like capital gains which are transfer payments have no bearing on the recorded national income , true or otherwise because such transfer payments do not enter national income computation.
→This report misses out an important aspect of estimating black income that is under-invoicing of sale of goods and over-invoicing of purchase of inputs. These are not under price and distribution control and hence are not ‘illegal’. Companies maintain separate account and use such incomes for non- sanctioned corporate purposes.
→The concept of black money is restricted to concept of black wealth. With the increasing service sector and improper tools to tap the black transactions happening in large scale, there is certain amount of black money associated with black incomes and turnover happening on account of transaction demand for money. This study rejects measures like demonetization which facilitate black production and black trading associated with current transactions.
→This report undermines the utility and validity of large number of black transactions and resulting black incomes connected with the turnover of the corporate sector. Professor Kabra argues that ‘if black incomes are estimated as a certain proportion of public expenditure, then it is equally necessary to estimate black incomes as a proportion of corporate sector turnover. This is because corporate wage bill permits reckonable diversion to black economy because corporate sector turnover is very large while the case is much less with salaried component of public spending. The former diversion is expected to be larger than the latter because there is little external control and the controller themselves are beneficiaries from falsification of accounts in the corporate sector.
→ A related point is the methods used for estimating the black income has a static framework, thus does not deal with the dynamics of the black economy. The uses of black incomes in consumption especially luxury and conspicuous consumption increases demand for real estate and other physical assets. This increases the rate of growth in production, construction and services. A severe crack down of black economy would lead to loss of production in many such activities. Professor N.S. JAGANNATHAN, in this context talks about the nexus between politics and sugar industry in U.P. , where he says ‘ sugar industry is helping the electoral process alive by making contribution in the expenses which does not have any other alternative’.
CONCLUSION
Given the limitations of the methods used to measure black economy, does it means we lose all hope and let it function? Clearly no. The importance of black economy could not be ignored because it affects day to day life of common people in the micro perspective. At the macro perspective, leakages arising out of black economy leaves impact on fiscal policies, balance of payments, leads to capital flight etc, thus the burden of borrowing on government increases. Most important, it does not allow the government to tap necessary resources out of taxes which has socio- economic implications and people often end up evading taxes owing to weak deterrence, moral standards, economic controls, inflation and government spending. As professor Jagannathan has commented ‘it is best to let the tax people use their accumulated expertise to deal with technical issues……. They know all about under and over invoicing’- they will eventually catch up. It is to see how many actually agrees with this view!

Aspects of black economy , Report of NIPFP (1985);
Focus on methods and estimations

TERM PAPER- PUBLIC FINANCE
COURSE CODE- 524

SUBMITTED BY- SUROBHI MUKHERJEE M.A.ECONOMICS, FINAL SEMESTER

BIBLIOGRAPHY

* ASPECTS OF BLACK ECONOMY IN INDIA- REPORT OF NIPFP (1985) by ACHARYA S. & associates * BLACK MONEY- COMMENT ON THE NIPFP REPORT ON ASPECTS OF BLACK ECONOMY by K.N. KABRA & N.S. JAGANNATHAN, NEW DELHI * THE BLACK ECONOMY IN INDIA by ARUN KUMAR