RAVINDRA H DHOLAKIA

T

he history of economic growth after independence in India has been an active field of empirical research leading to debates on effectiveness of policy regimes. The latest contribution by Balakrishnan and Parameshwaran (henceforth B-P, 2007), published in EPW on July 14, applying the state of the art statis tical techniques to find breakdates in the time trend of the gross domestic product (GDP) endogenously, is a welcome addition to this literature. The purpose of this note is to consider a few findings of earlier studies (missed out by B-P 2007) and thereby add to our outstanding of Indian economic growth. Had B-P gone back to 1992 for selecting their references, they would have found at least three studies directly relevant to their methodology and findings. These appeared in the Indian Economic Journal [Ganesh Kumar 1992], Indian Economic Review [Dholakia R H and Dholakia B H, 1993] and EPW [Dholakia R H 1994]. The first one applied the state of the art statistical techniques (available at that time) to the Indian GDP data from 1950-51 to 1989-90 and found breakdates for three broad sectors – primary, secondary and tertiary – and overall GDP. The second examined sources of economic growth in Indian agricultural sector by sub-periods similar to the findings of B-P (2007). The third, again applying the state of the art statistical techniques, identified breakdates in 20 state economies of India by the three broad sectors to gain an insight into which region and what sector triggered acceleration in Indian economic growth.

Breakdates by Sectors

Quandt (1958) and modified by Moschos (1989) to identity the breakdates from the time series data on Indian GDP by three broad sectors. He endogenously considered two breakdates by splitting the time series. His findings are relevant for meaningful comparison with B-P (2007). He found the breakdate for overall GDP to be 1981-82 while B-P (2007) finds it to be 1978-79 with confidence interval of 1977-78 to 1979-80. Fairly close but significantly different!

Secondly, Ganesh (1992) found agriculture (or primary sector) to be the basic driver for acceleration since it experienced an accelerating break from the trend in 1980-81. His finding is broadly corroborated by B-P (2007) that it was agriculture that first experienced acceleration in growth, releasing resources to the other sectors to pick up subsequently. Dholakia and Dholakia (1993) investigated the growth experience of agriculture and by eyeballing the time series data identified 1966-67 and 1980-81 as two breakdates. Their detailed estimates of sources of growth in the sector during these sub-periods confirmed the breakdates considered by them. B-P’s (2007) exercise broadly corroborates the first breakdate of the mid-1960s but not the second one of 1980-81. Ganesh Kumar’s exercise corroborates the second breakdate 1980-81 but not the first one of the mid1960s. Dholakia and Dholakia’s (1993) detailed analysis of sources of growth revealed the crucial role played by the application of modern agricultural inputs in accelerating growth in the sector. On the other hand, B-P (2007) question the role of the green revolution in accelerating agricultural growth permanently, based on identification of the breakdate in the sector which coincides with a bumper crop due largely to an extremely good monsoon. Since the weather factor was duly corrected in Dholakia and Dholakia (1993) exercise, their conclusions about application of modern agricultural inputs playing a crucial role in accelerating agricultural growth in India is more acceptable.

The second finding by Ganesh Kumar is regarding the structural breaks in the

Table: Estimates of Breakdates for State Economies in India

Sr States Primary Sector Secondary Sector Tertiary Sector Whole Economy No 1 2 3 4 5

--1972-73(+) 1968-69(+) 1979-80(+) ---

-1979-80(+) 1979-80(+) 1979-80(+)

-1980-81(+) 1972-73(+) 1967-68(+) 1982-83(–) 1975-76(+) 1973-74(+) 1973-74(+)

-1981-82(+) ----1984-85(+) 1985-86(+)

1985-86(–) -1973-74(–) 1985-86(–) --1975-76(+) 1985-86(+)

-1972-73(–) 1972-73(–) 1972-73(–)

-1978-79(+) 1979-80(+) 1979-80(+)

--1984-85(+) 1972-73(+) 1977-78(–) 1969-70(+) 1979-80(+) 1977-78(–) 1965-66(–) -1970-71(–) 1967-68(–)

-1984-85(+) 1979-80(–) -

-1982-83(+) 1974-75(+) -

-1979-80(–) 1983-84(+) -

--1975-76(+) 1972-73(+)

1973-74(+) 1974-75(+) 1976-77(+) 1974-75(+) 1982-83(+) -1972-73(+) 1982-83(+) 1979-80(+) 1981-82(+) 1982-83(+) 1981-82(+)

Notes: (1) (+) indicates acceleration and (–) indicates deceleration. Ganesh Kumar (1992) applied switch-(2) No adjustments have been made for weather or public administration. ing regressions with the test suggested by Source: Summarised from different tables in Dholakia (1994).

Economic and Political Weekly August 25, 2007 secondary sector. He found 1964-65 and 1981-82 as the breakdates which fall within the confidence intervals of the first two breakdates found by B-P (2007). Thus, there is a complete agreement on these breakdates by both the exercises. However, the relationship of the second breakdate of the secondary sector with the breakdate for the economy as a whole found by the two studies creates confusion. For Ganesh (1992), the two dates coincide while for B-P (2007), the breakdate for the economy as a whole precedes the one for the secondary sector.1

The most interesting finding of B-P (2007) is that services led the acceleration in growth in India because most of the sub-sectors within the tertiary sector experienced a trend-break before the economy as a whole. On the other hand, Ganesh (1992) found the breakdate for the aggregate tertiary sector as 1982-83, which is after the breakdate for the economy as a whole. When he adjusted for the subsector, Public Administration and Defence (PAD), the breakdate became 1981-82 for the remaining tertiary sector and the breakdate for the economy excluding PAD turned out to be 1980-81.

Comparison of the studies by Ganesh Kumar and B-P (2007) raises a few questions about the data and methodology. Both the studies tried to find the breakdates endogenously from the data set using the state of the art statistical techniques available at the time. Moreover, both the studies used Indian GDP data since 195051 at constant prices though Ganesh used 1980-81 prices while B-P used 1993-94 prices. The change of the base year for real GDP estimates might have some impact on the results. This is because whenever the base year for GDP estimates is changed, the past time series gets revised by revising the sub-sector series limited to only a certain level of disaggregation. Even this revision is generally not fully satisfactory and minor variations in the year to year growth at sectoral level are often found. The optimisation-based statistical technique used by B-P (2007) may, therefore, be sensitive to such variations in identifying precise breakdates. This is all the more important because the base year for GDP in India has now changed to 1999-2000 and the “comparable” official estimates of the past time series will soon be available. The same exercise on those data may give different estimates of the breakdates.

The methodology used by B-P (2007) is so sensitive that adding or subtracting data may have an impact on the identification of breakdates even in the past. Thus, potentially there are three reasons for not treating as final the identification of past breakdates in India applying the methodology developed by Bai and Perron (1998, 2003): One, change of base year of the series; two, consideration of additional data points; and three, non-adjustment of the weather factor, so critical in India till recently.

I had investigated [Dholakia 1994] the regional origins of acceleration in the growth of the Indian economy by identifying breakdates in the 20 state

Economic and Political Weekly August 25, 2007

In the tertiary sector, all states except Arunachal Pradesh and Haryana experienced trend-break during the period. As many as 14 states experienced acceleration from 1972-73 (West Bengal, Bihar and Andhra Pradesh) to 1984-85 (Maharashtra and Himachal Pradesh). Thus acceleration in the tertiary sector is the most widespread in the country. However, the national breakdate in the sector does not match with any state. It is possible that some disaggregation of the sector into trade and transport and financial and other services would provide better insights.

If we disregard sectors and consider states as originating impulses for growth acceleration, we find that different states experienced acceleration in their growth at different points. Bihar and Andhra Pradesh experienced it in late 1960s; Punjab and Haryana probably in the mid-1960s; Maharashtra, Tripura, Gujarat and UP in the early 1970s; Assam and MP in the late 1970s resulting in the acceleration of the national growth in 1981-82; and West Bengal, Karnataka and Himachal Pradesh in the mid-1980s to sustain the momentum.

This analysis was based on the data over 1960-61 to 1989-90 for most of the states at 1980-81 prices. With more recent data, different base year and more recent metho dology for identification of multiple breakdates, some of the estimates of the breakdates presented here could change. If a panel data considering states, sectors and years are considered for identification of the national breakdates in time trends, the results could be very different from the ones reported in B-P (2007). Such an exercise would be certainly very useful for understanding the Indian growth history though the breakdates so identified are subject to change as and when the base year changes or additional observations are available or newer techniques of analysis are employed.

Email: rdholkia@iimahd.ernet.in

Notes

1 It may be pointed out here that B-P (2007) do not report the breakdates for the secondary sector as a whole but report them separately for each sub-sector. The aggregative secondary sector may have different breakdates given their methodology. The same thing applies to tertiary sector in their exercise.

2 In Arunachal Pradesh and Assam, the time series began from 1970-71; in Haryana and Punjab from 1965-66; and in Himachal Pradesh from 1967-68. For Madhya Pradesh the base year was 1970-71.

References

Bai, J and P Perron (1998): ‘Estimating and Testing Linear Models with Multiple Structural Changes’, Econometrica, 66: 47-78.

– (2003): ‘Computation and Analysis of Multiple Structural Change Models’, Journal of Applied Econometrics, 18: 1-22.

Balakrishnan, P and M Parameshwaran (2007): ‘Understanding Economic Growth in India: A Prerequisite’, Economic and Political Weekly, July 14, 2915-22.

Dholakia, Ravindra H and Bakul H Dholakia (1993): ‘Growth of Total Factor Productivity in Indian Agriculture’, Indian Economic Review, XXVII(1), 25-40.

Dholakia, Ravindra H (1994): ‘Spatial Dimension of Acceleration of Economic Growth in India’, Economic and Political Weekly, August 27, 2303-09.

Ganesh, Kumar N (1992): ‘Some Comments on the Debate on India’s Economic Growth in the 1980s’, Indian Economic Journal, 39(4), 102-111.

Moschos, D (1989): ‘Export Expansion, Growth and the Level of Economic Development’, Journal of Development Economics, 30(1), 93-102.

Quandt, R E (1958): ‘The Estimation of a Linear Regression System Obeying Two Separate Regimes’, Journal of the American Statistical Assocation, 53, 873-880.

Back Volumes

Back Volumes of Economic and Political Weekly from 1976 to 2006 are available in unbound form.

Write to: Circulation Department, Economic and Political Weekly Hitkari House, 284 Shahid Bhagat Singh Road, Mumbai 400 001.

Economic and Political Weekly August 25, 2007