A Study of the Elasticity of the West Malaysian Income Tax System, 1961-70
Author: Mr. Nurun N. Choudhry
Publication Date:
01 Jan 1975
eISBN:
9781451947458
Language:
English
Keywords:
SP; currency; personal income income tax; tax structure; higher-than-unit income elasticity; income tax tax assessment; tax-to-base elasticity; Corporate taxes; Effective tax rate
Download PDF (859 KB)

NURUN N. CHOUDHRY *

Abstract

NURUN N. CHOUDHRY *

NURUN N. CHOUDHRY *

MICHAEL E. EDO *

It is well known that in estimating the built-in elasticity of a tax, historical revenue series must be adjusted to eliminate the effects on revenue of discretionary tax measures during the period under review. If no such adjustment is made, one obtains the buoyancy of the tax, which reflects the growth in revenue resulting both from the automatic growth in the base caused by the increase in income and from discretionary tax changes.

The method of adjustment depends on the available data on changes and the type and frequency of those changes. If disaggregated data on tax brackets or effective tax rates and on the changing composition of the bases are available, it is possible to construct a constant rate-base series that represents the hypothetical yields under a system assumed to remain unchanged throughout the period. In the case of full data availability, the construction of such a series would be reduced to an arithmetical exercise involving merely the multiplication of the tax bracket rates of the reference year by the corresponding base values and summing up the products for each year. Most often, however, insufficient information is available on legal rates and bases, and the fiscal economist has to use effective rates for income classes or commodity groupings.

This paper attempts to apply the constant rate structure method, involving the use of effective rates, to income taxes in Malaysia in order to estimate the elasticity of these taxes in the period 1961-70. As historical revenue data extending back to 1961 are not available for Sabah and Sarawak, this study is confined to West Malaysia. Section I explains the methodology employed in estimating the elasticities. Section II reviews estimates of buoyancies and elasticities of income taxes. In particular, since the personal income tax structure was considerably modified by the enactment of a new income tax ordinance in 1967, an attempt is made to compare the elasticity of the post-1967 structure with that of the earlier structure. Section III compares the elasticities and buoyancies derived through the constant rate structure method with those estimated on the basis of a different data adjustment method—namely, the proportional adjustment method employed by Prest and Sahota.1 Whereas the constant rate structure method was applied to assessed income tax series, the proportional adjustment method had to be applied to actual collection series because the desired degree of disaggregation was available only for assessments. It should therefore be kept in mind that the two sets of elasticities are derived from different data.

The exercise is carried out separately for personal and company income taxes, since companies are subject to an independent flat rate tax in Malaysia and dividends are taxable in the hands of taxpayers.

I. Methodology

personal income tax

The Annual Reports of the Department of Inland Revenue, Government of Malaysia, give the breakdown of assessed income and tax thereon by selected income groups—that is, assessees falling within selected income ranges.2 On the basis of this information, the effective rates of tax on each income group can be calculated for the reference year, and the application of these rates to the corresponding income groups in all the years generates a constant rate structure series.2 Since the income figures used relate to assessed income rather than to chargeable income, any changes in the law affecting the legal base (such as changes in personal allowances) have no effect on the reported base figures, and hence the adjustment must be made only for any changes in the rates.

The methodology of constructing the constant rate series may be expressed as follows:

TiP(t) = the assessed personal income tax, and Yi(t) = the assessed income in the ith income group

Tp(t)=Σi=1kTip(t) = the total assessed personal income tax in the period t

Y⁡(t)=Σi=1kTi(t) = the aggregate assessed income in the same year, and

r = the reference year.

Then, the average effective rate of taxation for the ith income group in the reference year is

ti(r)=Tip(r)Yi(r)

so that

Tp(r)=Σi=1kti(r)Yi(r)

Thus, the simulated assessed personal income tax in the tth year is

T^p(t)=Σi=1kti(r)Yi(t),t=1…n.

In Malaysia, the tax assessment for the current year is based on the income earned in the previous year—that is, the assessed income of year (t)—namely, Y(t) is the income earned in year (t-1). Hence, in computing the income elasticity of the income tax, it would be proper to relate the tax of a given year to the national income of the previous year.

For the sake of convenience in projection, and if the period covered is relatively short, it is usual to estimate a constant elasticity, using a double logarithmic function. Thus, the income elasticity of personal income tax can be derived as the slope of the regression equation

logT^p(t)=a+λlogXt−1(1)

where Xt-1 is a one-period lagged national income measure or any other suitable one-period lagged proxy. For estimating the elasticity of tax with respect to assessed income, T^p(t) is to be related to Y(t), since they belong to the same period, so that Xt-1 = Y(t).

For the purpose of this paper, three different aggregates—namely, gross domestic product (GDP) at current factor cost, factor income from labor and capital, and aggregate assessed income—have been chosen to make alternative estimates of the associated tax elasticities.

company income tax

No change took place in the basic rate of company tax in the sample period, except for a levy on tin profits imposed in 1966 and a 5 per cent development tax imposed in 1967. Because data on assessed tax classified by types of business were not available, effective rates for companies engaged in particular businesses could not be calculated. However, the same procedure as for the personal income tax was employed to construct a constant rate company income tax series on the basis of the reference year (r). Thus,

t⁡(r)=Tb(r)Yb(r)

and

T^b(t)=t⁡(r)Yb(t)

where T^b(t) is the simulated company income tax, Tb(r) is the assessed company tax in the reference year, and Yb(t) is the assessed company income in the tth year.

Company income tax assessment in each year is made with respect to a “basis year,” which is generally the business accounting year preceding the year of assessment. Thus, a considerable lag occurs in the assessment of company income; hence, in computing the elasticity of assessed tax to a proxy base or GDP, these variables should be lagged by at least a year.

Company income tax elasticity is obtained by employing the same form of the regression equation as for the personal tax. In addition to using lagged GDP, gross exports are also used as an independent variable with a one-year lag. In Malaysia, exports are nearly 45 per cent of gross domestic product. Rubber and tin account for a substantial fraction of company taxes, and these commodities represent nearly 50 per cent of exports. It is therefore appropriate to consider exports as a determinant of business taxes.

II. Estimated Buoyancy and Elasticity of Assessed Income Taxes

Buoyancy and elasticity of assessed personal and company income taxes are estimated separately and in combination from the observations for 1961 to 1969. Elasticities were computed for the 1961 and 1969 structures. The results for different explanatory variables are presented in Table 1.3

Table 1.

West Malaysia: Estimates of Buoyancy and Elasticity of Assessed Income Taxes1

article image

Statistical details are given in the Appendix, Table 12.

GDP = gross domestic product of West Malaysia (current prices), lagged by one year

FY = factor income from labor and capital, lagged by one year

YTA = total assessed income

YP = personal assessed income

X = gross exports, lagged by one year.

The overall buoyancy of income taxation and the buoyancies of the constituent parts (personal and company income taxes) are quite high. The buoyancy of total assessed income tax with respect to GDP is about 1.6; that of assessed personal income tax is 2.5 and is substantially higher than the buoyancy of 1.3 of assessed company taxes. This picture is little changed in a relative sense for elasticity estimates. The overall elasticity of the income tax is 1.1 under the 1969 tax structure, while the elasticity of personal income tax is about 1.7 and is again substantially higher than the magnitude of company income tax elasticity.

Estimates of buoyancies and elasticities of assessed income taxes with respect to other income aggregates enable us to check the tax-to-GDP buoyancy and elasticity through the partitioning approach. Consider the assessed personal income tax elasticity of the 1969 structure (1.66) with respect to GDP. According to the partitioning approach, this is explained by the product of the tax-to-base elasticity and base-to-GDP elasticity. The tax-to-base elasticity of assessed personal income tax is estimated to be 1.08, while that of assessed personal income to GDP is found to be 1.52 (Appendix, Table 12). The product of the two is 1.64, which is very close to the 1.66 value reported above. The elasticity of the 1961 personal income tax structure is only slightly higher at 1.11. Similarly, the buoyancy of 2.50 of the assessed personal income tax with respect to GDP can also be partitioned into the product of tax-to-base (1.62) and base-to-GDP (1.52) elasticities. These results suggest that the progressivity of the personal income tax contributed very little to raising the elasticity of the tax (that is, the tax has been almost proportional to its base); the higher-than-unit income elasticity is explained almost solely by the income elasticity of the base.

Although no further use is made here of the estimated elasticities and buoyancies of assessed income tax with respect to selected macroaggregates other than GDP, it is apparent that the partitioning approach may be quite useful to planners who are interested in altering some of the tax-to-base and/or base-to-GDP elasticities. In fact, in planning exercises, alternative estimates of tax revenues may be generated by assuming alternative planned values of some of these elasticities.

Several other observations can be made on the basis of the information presented in Table 1: (1) despite a substantial change in the rates of personal income tax introduced in 1967, there is little difference when the 1961 constant structure elasticities are compared with those of the 1969 constant structure; (2) buoyancies are generally higher than elasticities, reflecting the fact that the 1967 tax changes generated additional tax revenues; and (3) buoyancies and elasticities of the personal income tax exceed those of the company income tax. Finally, the income elasticities of business income tax are shown to be less than unity; since with a proportional tax, the tax-to-base elasticity is bound to be unity, this result for company tax-to-GDP elasticity indicates a lower-than-unity elasticity for assessed company income to GDP.

In 1967, the rate structure of the personal income tax, as embodied in the income tax ordinance of 1947, was substantially modified. It is interesting to inquire how much the ordinance of 1967 altered the pro-gressivity of the rate structure of the personal income tax. It raised the average statutory income rates without any change in the existing personal allowances and deductions. A comparison of the average statutory rates of 1961 with those of 1969 shows that these progressively increase up to the income level of M$30,000 (Table 2, column 3). Despite the increase in progressivity of the tax rates, the estimated elasticity of the personal taxes was slightly lower in 1969 than in 1961 (Table 1). This was caused by a shift in the distribution of assessed income in favor of the higher income groups, where the rate increases were relatively less progressive.

Table 2.

West Malaysia: Changes in Average Statutory Personal Income Tax Rates Between 1961 and 1969

(In per cent)

article image
Source: Department of Inland Revenue, Annual Report, 1961-69.

The marginal rate of taxation for income exceeding M$100,000 annually was 45 per cent in 1961 but 50 per cent in 1969.

The effective rates of tax on assessed personal income by income groups are shown in the Appendix, Table 10, which indicates that the aggregate average effective tax rate on assessed personal income has exhibited a secular rise from 6.6 per cent in 1961 to 9.6 per cent in 1969. In addition, the effective tax rates for individual income groups generally rose after 1964, with the marked rise in 1967 being accounted for by the discretionary measure of that year. However, the increase in the rates for individual income groups does not correspond favorably with the increase in average statutory rates. To cite an example, for the income group of M$20,001 to M$30,000 annually, the effective tax rate in 1969 was 12.5 per cent as against 9.8 per cent in 1961 (Appendix, Table 10). However, the average statutory tax rate for the same income group in 1969 ranged from 16.0 to 19.8 per cent as opposed to a range from 12.8 to 15.0 per cent in 1961 (Table 2). This difference between the increase in effective tax rate and that in statutory rate is evidenced for other income groups as well, although in varying degrees.

Two possible explanations may be offered for such discrepancies. First, the statutory tax rates apply to taxable or “chargeable” income which is arrived at after various exemptions and deductions are made from assessed personal income.4 This implies that an individual may be in a higher income group in terms of his assessed income but in a lower income group for tax purposes in terms of his taxable income. Secondly, changes in intragroup distribution of assessed and taxable income can give rise to a discrepancy between statutory and effective tax rates and also cause variations in effective tax rates for the assessed income groups concerned. Table 3 shows the combined effect of the above two factors in terms of the percentage changes in statutory and effective tax rates between 1961 and 1969. An examination of the table shows that while increases in statutory tax rates after the 1967 income tax legislation were progressive up to M$30,000 income annually, the percentage increases in 1969 over the 1961 effective rates were progressive across the board for personal income over M$ 10,000 annually. The percentage increases in the effective tax rates seem to suggest that a shift has occurred in intragroup distribution of personal income toward the top of the brackets between the two years.

Table 3.

West Malaysia: Percentage Change in Average Statutory and Effective Income Tax Rates Between 1961 and 1969

article image
Source: Department of Inland Revenue, Annual Report, 1961-69.

Lack of data has made it impossible, to obtain a distribution of taxable personal income by income groups. Fortunately, data on distribution of assessed personal income by income groups were available. The distribution of personal assessed income by income groups on a cumulative basis is shown in the Appendix, Table 11, which indicates a slight secular rise in the cumulative distribution of personal income in the higher income groups.5 Thus, the share of assessed personal income up to M$15,000 annually fell from 65 per cent in 1961 to 62 per cent in 1969. Similarly, the cumulative distribution of assessed personal income up to M$30,000 annually recorded a decline from 86 per cent in 1961 to 82 per cent in 1969. But, since the redistribution was slight, the constant rate of tax-to-base elasticities for 1961 and 1969 is just above unity. The finding that the elasticity of the 1969 structure is only slightly lower than that of the 1961 structure is explained by the fact that the lessened progressivity (caused by the 1967 change) in the upper income ranges was compensated for by the increase in progressivity in the lower ranges.

Since the decline in the elasticity of the personal income tax structure was very small, the overall elasticity and the buoyancy of the system remain relatively high. The high buoyancy of the personal income tax reflects several factors. First, the elasticity of personal income to GDP is quite high. Secondly, the effective tax rate per income group has increased—in part because of the progressive rise in the statutory rates and in part because of the intragroup shifts in income distribution toward the top of the income group. Finally, the intergroup distribution of income has shifted toward the higher income groups.

While the buoyancy of company taxes with respect to GDP is greater than unity (1.27), the constant structure elasticity is less than unity (0.97). Prior to 1967, the company income was taxed at a flat rate of 40 per cent. As already noted, the 1967 income tax legislation imposed a 5 per cent development surcharge on all company income.6 As far as the constant structure elasticity is concerned, the 5 per cent surcharge cannot make any difference because the company income tax is not progressive. Consequently, the elasticity of the company income tax with respect to GDP is the same for both the 1961 and 1969 tax rates. While the absence of progressivity in the company tax rate implies that tax-to-base elasticity must be unity, if the elasticity of the tax with respect to GDP is less than unity, the elasticity of company income with respect to GDP must be less than unity and equal to the income elasticity of the tax on it. This appears to be borne out by the computations, which yielded the same value of 0.97 for the elasticity of company income with respect to GDP. The buoyancy of assessed business income taxes with respect to GDP was greater than unity because of the 5 per cent development surcharge levied in 1967.

West Malaysia exports nearly 45 per cent of its GDP, mostly in primary exports or related manufactured products. Fluctuations in the world demand for Malaysia’s exports will therefore have a significant effect on company income and hence on tax receipts. The estimated elasticity of 1.24 of assessed company tax with respect to gross exports (lagged by a year) seems to confirm the importance of export fluctuations for company tax receipts.

The above discussion on the behavior of the elasticity of personal and company taxes suggests that the income tax ordinance of 1967 did not materially change the overall elasticity of the income tax system. The tax-to-base elasticity is not very different from unity because of the relative stability of the distribution of income in the tax brackets.

III. Buoyancy and Elasticity of Collected Income Taxes

Buoyancy and elasticity of actual income tax collections have been estimated in order to compare them with assessed income taxes. The elasticities of collected income taxes have been estimated on the basis of the proportional adjustment method (the Sahota-Prest method), since disaggregated data on income tax collections were not available for the constant structure method. These comparisons are shown in Table 4.

Table 4.

West Malaysia: Buoyancies and Elasticities of Assessed and Collected Income Taxes1

article image

All buoyancies and elasticities are estimated with respect to gross domestic product.

As expected, the estimated buoyancies of total income tax collections and of personal and company tax collections are higher than the corresponding elasticity estimates. The magnitudes of these elasticities are quite high, however. Interestingly enough, although personal income tax collection is relatively more buoyant than company income tax collection, its elasticity is lower than that of the latter.

A comparison of the buoyancies and the elasticities of actual income tax collections with those of income taxes assessed reveals the following: (1) buoyancies and elasticities of actual total and company income tax collections are higher than the corresponding buoyancies and elasticities of assessments; (2) both the buoyancy and the elasticity of actual personal income tax collection are lower than the corresponding estimates for personal income tax assessment; and (3) the differences in buoyancy and elasticity between personal and company income tax collections are lower than those between personal and company income assessments. These observed differences between buoyancies and elasticities of actual and assessed income taxes imply differences in their growth rates; actual total taxes and company income taxes grew at a higher rate, on an average, than assessed total taxes and company taxes, while actual personal income tax collection increased at a lower average rate than assessed personal income tax. Trends in the ratio of assessments to collections are shown in Table 5.

Table 5.

West Malaysia: Ratio of Assessed to Collected Income Taxes, 1960-70

article image

During most of the 1960s, total income tax collections fell short of assessments. The ratio of total income tax assessments to total income tax collections, however, declined over the years, despite considerable fluctuation in some years. This decline implies a relatively high average growth rate of actual total income tax collection as compared to total income tax assessment and is reflected by its estimated buoyancy and elasticity, both being greater in magnitude than those of total income tax assessment. The case of company income tax collections was similar; although company income tax collection fell short of its assessment, it grew at a higher average rate than did the latter. Personal income tax collection, on the other hand, grew at a slightly slower pace than personal income tax assessment.

The higher magnitudes of the elasticities of income tax collection than those of income tax assessment appear to indicate that income tax administration was improving over time. However, the smaller difference between the elasticities of personal income tax collection and of its assessment appears to suggest that the improvement mostly took place with respect to company tax administration.

APPENDIX

Table 6.

West Malaysia: Gross Domestic Product and Exports, 1960-70

(In millions of Malaysian dollars)

article image
Sources: Department of Statistics, Kira2Negara Bagi Malaysia Barat (National Accounts of West Malaysia), 1960-68 (1972); Bank Negara Malaysia, Quarterly Bulletin, Vol. 5 (December 1972), Table 7, Column 3, p. 63.

Includes re-exports.

Table 7.

West Malaysia: Private Income from Labor and Capital at Current Prices, 1960-70

(In millions of Malaysian dollars)

article image
Sources: Department of Statistics, Kira2Negara Bagi Malaysia Barat (National Accounts of West Malaysia), 1960-68 (1972), Table 4, p. 10.

Here gross domestic product refers to Table 6, column (1).

From 1969 to 1970, net labor and capital income have been estimated by assuming the trend ratio to be 0.77 and multiplying it by gross domestic product as shown in Table 6, column (1).

Table 8.

West Malaysia: Assessed Income Tax, 1960-70

(In millions of Malaysian dollars)

article image
Source: Department of Inland Revenue, Annual Report.
Table 9.

West Malaysia: Assessed Personal Income and Company Income, 1960-70

(In millions of Malaysian dollars)

article image
Source: Department of Inland Revenue, Annual Report.
Table 10.

West Malaysia: Personal Income Taxation Rates on Assessed Personal Income by Groups, 1960-69

(In per cent)

article image
Source: Department of Inland Revenue, Annual Report.
Table 11.

West Malaysia: Cumulative Distribution of Assessed Personal Income by Income Groups, 1960-69

(In per cent)

article image
Source: Department of Inland Revenue, Annual Report.
Table 12.

West Malaysia: Estimates of Assessed Income Tax Equation

article image

GDP = West Malaysian gross domestic product at current factor cost.

YTA = total assessed income (personal and business)

YP = personal assessed income

FY = factor income from labor and capital

X = gross exports.

Serially correlated.

*

Mr. Choudhry, an economist in the Fiscal Analysis Division of the Fiscal Affairs Department, holds a master’s degree in statistics from the University of Dacca (Bangladesh) and a doctorate in economics from the University of California at Berkeley.

1

For a description and evaluation of this method, see Roy W. Bahl, “Alternative Methods for Tax Revenue Forecasting in Developing Countries: A Conceptual Analysis” (unpublished, International Monetary Fund, October 16, 1972).

2

See the Appendix, Tables 10 and 11.

3

The statistical results of the regression equations are presented in the Appendix, Table 12.

4

Assessed personal income is not taxable income, because the former constitutes income before the granting of allowances and other deductions, which are fairly substantial in Malaysia. Data on taxable income by income group were not available. However, it has been observed from aggregate data that taxable personal income was around 54 per cent of assessed personal income.

5

This only shows that higher assessed incomes account for a larger share of total assessed income and does not necessarily imply an increase in the equality of incomes among families.

6

In 1966 a tin profit tax levy was imposed on personal income at the rate of 10 per cent in excess of M$ 10,000 of tin profit income, but it was not to exceed the amount by which tin profit exceeded M$ 10,000.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Papers
IMF Staff papers: Volume 22 No. 2
Author: International Monetary Fund. Research Dept.
Volume/Issue:
Volume 1975: Issue 002
Publisher:
International Monetary Fund
ISBN:
9781451947458
ISSN:
1020-7635
Pages:
338
DOI:
https://doi.org/10.5089/9781451947458.024