Sectoral volatility, development, and governance: a case study of Pakistan.

Azid, Toseef ; Khaliq, Naeem ; Jamil, Muhammad 等

I. INTRODUCTION

Development of overall economy of any country largely depends upon the characteristics of different prominent sectors such as agriculture, industry, services, etc. Sharp structural change in prominent sectors are experienced by the Pakistan's economy during the last four decades, in which industrial and service sector have exhibited an extra ordinary rate of growth, while the agricultural sector did not shown that rate of growth which was experienced during the time of green revolution. Due to these structural changes in the prominent sectors volatility of growth rate has been experienced by the economy.

To the extent that most of the recent volatility in growth rate of GDP can be attributed to the increasing share of the some volatility of the some prominent sectors, the analysis of their volatility can be useful in providing some enlightenment on the factors behind this phenomenon and its implications for the formulation of the policy in the future.

The main objective of this study is to use a time series analysis to analyse the actual cause of the volatility in the output/growth rate of output of the Pakistan's economy. The technical characteristics of the volatility of the different sectors will be analysed and then an effort will be made to estimate the impact of sectoral volatility on the volatility of the growth rate of the Pakistan's economy. Specific questions which will be addressed in the course of study will include: What are the main characteristics of the structure of the Pakistan's economy? What is the nature of volatility of the different economic sectors? Does sectoral volatility explains relative changes in the growth rate? In other words, is sectoral volatility biased or neutral? How the volatility in the different sectors of Pakistan's economy is correlated with each other? To what extent do volatility in growth rate is associated with the Volatility of the growth rates of the sectors under analysis? What are the main implications of the volatility parameters for the Pakistan's policy problem, and for the achievement of stable growth rate? Does instability in political structure affect volatility of growth rate of output?

This will represent the first attempt to analyse quantitatively the relationship between the volatility and development. The study is organised in the following manner. Following the section of Introduction Section II presents the literature review while Section III presents methodology and estimations of impact of sectoral volatility on the performance of Pakistan's economy. Section IV discusses data and construction of variables and results and there discussion will be presented in Section V. whereas last section presents the summary and conclusion along with some policy suggestions.

II. LITERATURE RIVIEW

In the economic literature, a number of efforts can be seen discussing the impact of volatility on the economic performance of different countries especially the developing countries, e.g., Koren and Tenreyro (2005) explained that despite the number of steps have been taken by most of the developing countries towards the stability of their economies, still one can easily observe the volatility in most of their macroeconomic variables. The concern with volatility in most of the developing countries arises day by day. Most of the studies are concerned with this question: why GDP growth is so much volatile in poor countries than in rich one'? Generally experts identified the four possible reasons:

(a) Poor countries specialise in more volatile sectors;

(b) Poor countries specialise in fewer sectors;

(c) Poor countries experience more frequent and more severe aggregate shocks (e.g. from macroeconomic policy);

(d) Poor countries' macroeconomic fluctuations are more highly correlated with the shocks of the sectors they specialise in.

This is the requirement of the time that how to decompose volatility into these four sources, quantify their contribution to aggregate volatility, and study how they relate to the stage of development.

However, a number of studies can be seen in the literature discussing the phenomenon of volatility and its impact on the performance of the economy, It is well recognised that volatility of different sectors has negative impact on the performance of the economy. Especially in the literature, it has been observed that volatility of those sectors in which the economy is specialised has a significant effect on the production and trade of the developing as well as the developed economies. For example, macro economic impact of volatility discussed by Koren and Tenreyro (2005), Lucas (1988), Acemoglu and Zilibotti (1997), Obstfeld (1994), Saint-Paul (1992), Greenwood and Jovanovic (1990), Imbs and Wacziarg (2003), Stockman (1988), Scheinkman and Xiong (2003), Cheema (2004), Perotti (1996), Atkinson (1996 1997), Gottschalk and Smeeding (1997), Bourguignon and Morrissson (1998), Li, et al. (1998), Betancourt (1996) Mobarak (2001), Pritchett (2000), Jalan and Ravallion (1999) and Morduch (1995), Lucas (1987), Pallage, et al. (2003), Wolfers (2003), Barlevy (2002), Barro and Sala-I-Martin (1995) and many others. Whereas determinants of Volatility are discussed by Levine and Renelt (1992), Acemoglu and Zilibotti (1997), Rodrik (1999) and Ramey and Valerie (1995). Relationship between democracy and volatility explained by Henisz (2000), Nooruddin (2003). Chandra (1998) and Quinn and Woolley (2001).

In spite of the crucial importance attributed to growth rate of GDP in Pakistan like other developing countries, no empirical quantitative research has however, been conducted to examine the volatility of that sector in which economy is specialising, and its impact on the volatility of growth rate in Pakistan. The present study focuses on this issue. In the first step a general overview regarding the historical patron of economic volatility in Pakistan is given in the following section. (1)

II.I. Patron of Economic Volatility in Pakistan

Since independence the economy of Pakistan has undergone dramatic structural changes and economic growth. Pakistan has tried to change its economic structure as the other underdeveloped countries from an agricultural economy to an industrial exportoriented economy, in which the manufacturing sector constitutes today the dominant form of economic activity. This dominance resulted from a development strategy based on tax exemption schemes, in addition to other incentives, which the Pakistan's government implemented in the past five decades. The main objective of different schemes was to alleviate the historically high level of unemployment and at the same time, promote the economic and social welfare of the population.

During the last five decades the manufacturing sector itself has experienced a series of changes in its internal structure. The structure is based towards the capital intensive techniques instead of labour-intensive one. This however, created the problem of unemployment, balance of payments (most of the intermediate and final inputs are imported). The high technology is also attracted by the government and main industrial groups in Pakistan. One of the main reasons for the attraction has been the necessity to maintain or improve the international competitiveness of Pakistan's manufacturing sector. The imported technology used by these industries and the associated technical changes have also affected the utilisation of layout and has contributed towards increase in labour productivity. It has induced changes in the organisation and composition of the work force, and affected skill requirements and management of labour. Despite all of this, still agriculture plays a significant role in the development of Pakistan's economy.

The pattern of economic volatility in Pakistan is complex. At the macroeconomic level the very high volatility recorded in real growth rates, price inflation, and private investment per capita, government revenues per capita, terms of trade and real exchange rate. But patterns of volatility vary among sectors. In terms of GDP the most volatile sectors are agricultural, industrial and service; while the least volatile are distribution, transport, and communications. On government expenditure current expenditures (there are three major components of current expenditure, namely, interest payments, defense and expenditure on civil administration) are highest than the development expenditures while public expenditure as percentage share of GDP has been relatively stable.

Tax and debt funded public spending as the driving force of the Pakistan's economy. Fiscal policy and budget management constitute the pivot of macro-economic policy. Major problems include: excessive centralisation of resources and powers, to the detriment of sub-national units of government; prevalence of fiscal imbalances both vertically horizontally; and frequent overlapping and non-coordination of expenditure responsibilities among different levels of government. There is need to evaluate and restructure present fiscal set-up, to ensure fiscal discipline at all levels, as well as to secure greater understanding and cooperation across the different tiers of government.

Exchange rate policy is a key factor in economic management. In an elusive search for a real exchange rate to maintain both internal and external balance, Pakistan has experimented with a succession of exchange rate regimes. The latest experiment is based on managed floating rate of exchange.

Since its beginning in 1947 national development planning in Pakistan has suffered from lack of systematic, integrated and target oriented approach, each plan being essentially a laundry list of projects, some rolled over from the over from the previous ones. Lack of clear vision, transparency and functional cooperation at the political levels, marginalisation of civil society in the planning process, and lack of rigor at the bureaucratic level have severely compromised the quality of the planning. At all levels, of government technical expertise as well as technology and information management systems are very deficient.

The review of policy options considers short-middle term as well as long term prospects. It focuses on monetary policies, prices and exchange rate management, revenue stabilisation, diversification and growth, public expenditure management and the constitutional and operational problems of fiscal federalism. The main determinant of the stock of money, in Pakistan, has been the consolidated fiscal balance of all levels of government, federal, state and local, which has been in deficit for most of the time since 1947. The money supply growth has contributed to the relative growth of the service sector and the relative decline of the agricultural sector of the economy, contributing to considerable GDP volatility. The federal government and most state governments have embarked on programs to improve public expenditure management by downsizing, rightsizing and restructuring the public services and privatising public enterprises the stabilisation of public expenditure is constrained by the lack of harmonisation and coordination of expenditure management by the various tiers of government

Public revenue in Pakistan is inadequate and unstable. The major cause of revenue volatility is a combination of two factors: the large and unpredictable fluctuations in agricultural sector because the whole economic activity is based on agriculture and ad hoc policies as well as inefficient structure of the central board of revenue. In the short run efforts should be made to raise more revenue through more effective harvesting of existing sources and more imaginative investigation and development of new ones. In the long run steps should be taken to promote and support increased production and productivity in the various sectors of the economy. The low level of social development and social security is a major constraint to sustainable growth. There is need to enlarge the revenue base through social security taxation in order to provide adequately for the necessary investment in social service delivery.

Most studies on the volatility structure and development have been undertaken for developed and developing countries as will be reviewed in the study. Only a very limited number of studies deal with these issues in less developed countries (LDCs). Still no serious attempt has been seen covering the area of development and sectoral volatility. This is the first systematic quantitative study on the measurement of volatility and development. In addition, it is the first attempt to provide a quarterly time series data set covering the period 1971-72 to 2002-2003, which capture the different shocks of the Pakistan's economy and adjustments associated with the different economic and political crises. The analysis of the relationship between the sectoral volatility and growth rates in Pakistan makes it a unique study in views of future policy options.

III. DATA AND METHODOLOGY

III.I. Methodological Issues

The study takes advantage of the developments in the theory of unit root test, Vector Auto Regressive Model (VAR), Co-integration Test and Impulse Response Functions (IRFs). To measure the time varying measure of volatility of output, economists construct a rolling (moving) variance of the series. However, the rolling variance is a naive derivation of uncertainty because economic agents are not necessarily exploiting patterns in the data when making forecasts of uncertainty through measures of fluctuations but not of uncertainty. The choice stands for a measure of uncertainty measure obtained through the ARCH-GARCH process. Auto-regressive Conditional Heteroscedasticity (ARCH) models were introduced by Engel (1982) and generalised auto-regressive models (GARCH) by Bollerslev (1986).

(1) Autoregressive describe a feedback mechanism that incorporates past observations into the present.

(2) Conditional implies a dependence on implies a dependence on the observations of immediate past.

(3) Heteroscedasticity represents a time-varying variance (i.e. volatility).

Therefore ARCH models allow the error term to have a time varying variance i.e. to be conditional on the past behaviour of the series. In the present study volatility of all the variables is calculated using rolling (moving) standard deviations of the series and ARCH-GARCH process. Under the rolling (moving) standard deviation as the measure of volatility 4-quarter moving standard deviation and 8-quarter moving standard deviation are used for analysis.

A dummy variable is used to check the impact of political stability on the volatility of output in growth and level form. Value of dummy variable is one for the periods of election campaign (one quarter before, during and after the government change) and zero otherwise. It is expected that political instability lead to high volatility in output in growth and level form.

III.II. Data and Variables Notations

The study uses data of output (GDP), value added of agriculture, value added of Finance and Insurance, value added of services, value added of industry and value added of whole sale and retail. The secondary quarterly data covering the time period 1971-72 to 2002-2003 is used that has been taken from Kemal and Arby (2004). (2)

To differentiate among different types volatility variables that are calculated using moving standard deviations of the series and ARCH-GARCH process following notations are used:

To differentiate among different types volatility variables that
are calculated using moving standard deviations of the series
and ARCH-GARCH process following notations are used:

 4 quarter moving Standard deviation =VOL
 8 quarter moving standard deviation =VOLL
 Volatility based on ARCH-GARCII =VOLT

Variables
 Output (GDP) =Y
 Value added of agriculture =VAG
 Value added of Finance and Insurance =VFIN
 Value added of Services =VSER
 Value added of Industry =VIN
 Value added of Whole sale and retail =VWH

Growth Rate of Variables
 Growth rate of output =GRY
 Growth rate of value added of agriculture =GR VAG
 Growth rate of value added of Finance and Insurance =GR VFIN
 Growth rate of value added of Services =GR VSER
 Growth rate of value added of Industry =GR VIN
 Growth rate of value added of Whole sale and retail =GR VWII

Volatility based on 4 quarter moving standard deviation
 Volatility of output =VOL Y
 Volatility of value added of agriculture =VOL VAG
 Volatility of value added of Finance and Insurance =VOL VFIN
 Volatility of value added of Services =VOL VSER
 Volatility of value added of Industry =VOL VIN
 Volatility of value added of Whole sale and retail =VOL VWH

Volatility based on 8 quarter moving standard deviation
 Volatility of output =VOLL_Y
 Volatility of value added of agriculture =VOLL_VAG
 Volatility of value added of Finance and Insurance =VOLL_VFIN
 Volatility of value added of Services =VOLL_VSER
 Volatility of value added of Industry =VOLL_VIN
 Volatility of value added of Whole sale and retail =VOLL_VWH
Volatility based on ARCH-GARCH Process
 Volatility of Growth rate of output =VOLT_GRY
 Volatility of Growth rate of value added of =VOLT_GR_VAG
 agriculture
 Volatility of Growth rate of value added of Finance =VOLT_GR_VFIN
 and Insurance
 Volatility of Growth rate of value added of Services =VOLT_GR_VSER
 Volatility of Growth rate of value added of Industry =VOLT_GR_VIN
 Volatility of Growth rate of value added of Whole =VOLT_GR_VWH
 sale and retail

V. RESULTS AND DISCUSSION

The data for this study exhibits the regular characteristics associated with most of the macroeconomic variables. This conclusion derives by looking at various tests carried out on the variables used. Simple graphical comparisons of volatility variables obtained through the moving standard deviation (both 4-quarter and 8-quarter) indicate that volatility of output is the highest volatile sector followed by volatility of value added of agriculture sector. Dispersion among the volatilities of other variables is relatively less. Important thing to note down is this that all the volatility variables whether based on 4-quarter or on 8-quarter moving standard deviation are increasing over the time period. This type of patron can be seen from Figure 1 and Figure 2.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

Volatility of growth rate of value added of finance and insurance is observed to be most volatile when observing the volatility variables through ARCH-GARCH process (3). Volatility of other growth variables are less volatile and lie within [+ or -] 2. Here the important thing is that volatility variables obtain through ARCH-GARCH process can have negative values.

Now moving towards more sophisticated techniques of analysis in section V.I results of unit root tests are reported, this is necessary because the co-integration tests can be applied only to variables that are non-stationary in levels (contain unit root). In section V.II error correction model and regression analysis are presented to check the short run or long run relationship between volatility of growth rate of output and variables under analysis and in section V.III Impulse Response Functions (IRF) will be presented.

V.I. Unit Root Tests

Checking stationary is necessary because during building models for time series, the underlying stochastic process that generated the series must be invariant with respect to time. If the characteristics of the stochastic process change over time, i.e., if the process is non-stationary, it will often be difficult to represent the time series over past and future intervals of time by a simple algebraic model. This leads to misleading result. On the other hand, if the stochastic process is fixed in time, i.e., if it is stationary, then one can model the process via an equation with fixed coefficients that can be estimated from past data.

We report the results for the Augmented Dickey Fuller (ADF) test that has been initially developed by Dickey and Fuller (1979) because it has an Over-riding advantage on other tests, as ADF automatically controls for higher order correlations by assuming that the coefficient of the series follows an AR (p) process and automatically adjusts the test methodology. Results of unit root tests on variables are reported in Table I.

Results of ADF tests shows that all the variables of the model are integrated of order one that is I (1) if test is applied with intercept, suggesting the need for differencing of the variables. Results of the unit root tests based on with trend and intercept or with none showed mixed level of integration that is variables are not integrated of same order. All the growth rate variables are integrated of order zero whether based on with intercept, with trend and intercept or with none.

Our prime objective is to check the impact of volatility of different sectors on the volatility of output. In this regard we applied unit root tests on the volatility of different variables that are calculated by 4-quarter moving standard deviation, 8-quarter moving standard deviation and by ARCH-GARCH process. Here the results of ADF test indicate that all the volatility variables of growth rate variables based on ARCH-GARCH process are integrated of order zero. Volatility variables based on moving standard deviation showed mixed results. Volatility variables based on 4-quarter moving standard deviation and 8-quarter moving standard deviation are integrated of order one except volatility of value added of finance and insurance and volatility of value added of services when test applied with intercept. Volatility of value added of services based on moving standard deviation (both 4-qaurter and 8-quarter) is integrated of order zero when ADF test applied with intercept or with trend and intercept while integrated of first order when test was applied with none. Volatility of value added of finance and insurance based on moving standard deviation (4-quarter moving and 8-quarter) is integrated of order zero when test was applied with intercept otherwise integrated of order one. According to Angel-Granger Approach if any of the variables is integrated of order zero then co-integration test cannot be applied. So there is no co-integration among volatility of growth rate of output and volatility of growth rates of value added by different sectors under analysis, all there exist is the short run relationship.

V.II. Error Correction Analysis and Regression Analysis

In this section analysis has been performed in two steps in the first step volatility of growth rate of value added of each variable is regressed over the volatility of growth rate of output. In the second step all the variables of volatility of growth rates of value added of different variables used to check the impact on volatility of growth rate of output at once. Another attempt is made to test the hypothesis based on the volatility derived from moving standard deviations. Dependent variable is volatility of the output (based on 4-quarter and 8quarter moving standard deviation) and independent variables are volatility of (based on 4-quarter and 8-quarter moving standard deviation) value added of different sectors under analysis. Results are provided in Table 2.

Results are very much in the same direction as was expected. From the regression results it has been observed that volatility of growth rate of selected sectors have significant impact on the volatility of growth rate of the income when regressed combined or separately. Similar results were observed in case of volatility variables obtained through moving standard deviations except of volatility of finance and insurance. Volatility of finance and insurance obtained through moving standard deviation has significant impact on the volatility of output when regressed separately while indicate negative but insignificant impact on the volatility of output when combined with other variables in regression.

In magnitude form volatility of growth rate value added of services contribute highest and volatility of growth rate of value added of finance and insurance contribute lowest to volatility of growth rate of output when regressed separately or combined with other variables. Results of volatility variables based on moving standard deviations (based on both 4-quarter and 8-quarter) indicate that volatility of value added of whole sale and retail contribute highest volatility of value added of agriculture contribute lowest to the volatility of output when regressed separately. When combined with other variables indicate that volatility of value added of services contributes highest and volatility of value added of industry contributes lowest to the volatility of output.

At the end dummy variable constructed for the political instability used as another independent variable. Results indicate that political instability has insignificant effect on the volatility of output in growth and level forms. As it is observed that volatility of financial sector do not have significant impact on the on the volatility of output. So a new regression is estimated without this variable and included the same dummy. However, the similar results were obtained that is political instability have no significant effect on the volatility of output.

One can also find the short run one by constructing an error correction mechanism (ECM). A pre-condition of the ECM is this that all the variables should be integrated of same order and no variable should be integrated of order zero. If any variable is integrated of order zero or integrated of different orders then there do not exist long run relationship so no adjustment process. This again put another brick in the wall of analysis that there exist short run relationship among the volatility of sectors under analysis and volatility of growth rate.

V.IlI. Impulse Response Functions (IRF)

The findings of Impulse Response functions are not very much promising. It has been observed from Figure 3 to Figure 7 that volatility of sectoral growth has not significant impact on the performance of the economy in the long-run. Figure 3 presents the impulse response function of volatility of growth rate of value added of agricultural sector to one standard deviation shock to volatility of growth rate of income and the IRFs indicate that impact is temporary. The volatility of growth rate of income gradually returns to the converging point. Previous literature does not suggest any a priori explanation of this behaviour. The effects of volatility of growth rate of finance and insurance sector are presented in Figure 4. Same phenomenon has been occurred as observed in case of volatility of growth rate of value added of agriculture. Short run fluctuations can be seen whereas long run effects are not appeared. Currently it is well known fact that increases in the volatility of finance and insurance sector has the impact in the performance of the economy but the impact is observed to be temporary. The impact of volatility of growth rate of industrial sector, volatility of growth rate of service sector and volatility of growth rate of whole sale and retail sector are not significantly different from zero.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

VI. CONCLUSIONS

From this study it has been observed that in Pakistan every policy is based on short-run whereas it requires the introduction of a long term planning and expenditure framework for Pakistan. This requires an appropriate policy and institutional framework which addresses the long-term goals of the nation as determined through a transparent process involving all the legitimate stakeholders, and should be based on a clear strategy and an integrated program of action. The National Economic Council has a central role to play in this task.

However, for this study it is also observed that volatility of different sectors have impact on the volatility of growth rates in the short-run while volatility of value added of finance and insurance indicate insignificant impact. Volatility of value added of services contributed highest to the volatility of growth rate of output and volatility of finance and insurance contributed lowest to the volatility of growth rate of output.

Dynamic changes in political structure have insignificant impact on the growth rate of the economy. Currently the insignificant impact might be due to the dummy variable used for the political instability. So there is a need to estimate the impact of political instability on the volatility of the growth rate by considering and developing comprehensive measures of political instability.

The problem of inadequate, untimely and unreliable data has adversely affected development planning and management. Although there are several institutions at the federal level charged with the production of statistical and survey reports, their performance has been uneven and irregular. The system lacks the capacity to harvest and use the information available at the various agencies and centres of action.

There is need to re-think and restructure fiscal federalism in Pakistan. Under prolonged military rule the principles and practice of fiscal federalism were eroded. Efforts are now being made to re-build the system. Such efforts should include the establishment of mechanisms for coordination and cooperation between Federal Government of Pakistan and provincial governments in such a way as to make it possible to agree on economy-wide macroeconomic objectives and targets, and ways of achieving same. The Constitution itself should be drastically reviewed and refashioned in the light of the needs and expressed wishes of the people.

REFERENCES

Acemoglu, Daron and Fabrizio Zilibotti (1997) Was Prometheus Unbound by Chance? Risk, Diversification, and Growth. Journal of Political Economy 105:4, 709-751.

Atkinson, A. B. (1996) Seeking to Explain the Distribution of Income. In John Hills (ed.) New Inequalities: The Changing Distribution of Income and Wealth in the United Kingdom. Cambridge: Cambridge University Press. 19-48.

Atkinson, A. B. (1997) Bringing the Income Distribution in from the Cold. Economic Journal 107:441,297-321.

Barlevy, Gadi (2002) The Cost of Business Cycles under Endogenous Growth. Northwestern University. (Mimeographed.)

Barro, Robert and Xavier Sala-I-Martin (1995) Economic Growth. New York: McGrawHill.

Betancourt, Roger (1996) Growth Capabilities and Development: Implications for Transition Processes in Cuba. Economic Development and Cultural Change 315-331.

Bollerslev, Tim (1986) Generalised Autoregressive Conditional Heteroskedasticity. Journal of Econometrics 31,307-327.

Bourguignon, F. and C. Morrisson (1998) Inequality and Development: The Role of Dualism. Journal of Development Economics 57, 233-57.

Chandra, Siddartha (1998) Plurality of Institutions and Economic Growth. University of Pittsburgh. (Mimeograhed.)

Cheema, Faisal (2004) Macroeconomic Stability of Pakistan: The Role of the IMF and Worm Bank (1997-2003). University of Illinois at Urbana-Champaign, Available at http://www.acdis.uiuc.edu/Research/ops/cheema/cover.html Access Date 26th March, 2006.

Dickey, D. A., and W. A. Fuller (1979) Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of the American Statistical Association 74, 427-431.

Gottschalk, Peter and Timothy M. Smeeding (1997) Cross-National Comparisons of Earnings and Income Inequality. Journal of Economic Literature 35,633-687.

Greenwood, J. and B. Jovanovic (1990) Financial Development, Growth, and the Distribution of Income. Journal of Political Economy 98:5, 1076-1107.

Henisz, Witold (2000) The Institutional Environment for Economic Growth. Economics and Politics 12: 1, 1-31.

Imbs, J. and R. Wacziarg (2003) Stages of Diversification. American Economic Review 93:1, 63-86.

Jalan, Jyotsna and Martin Ravallion (1999) Are the Poor Less Well Insured? Evidence on Vulnerability to Income Risk in Rural China. Journal of Development Economics 58, 61-81.

Kemal, A. R. and M. F. Arby (2004) Quaterisation of Annual GDP of Pakistan Pakistan Institute of Development Economics, Islamabad. (Pakistan Statistical Paper).

Koren, M. and S. Tenreyro (2005) Technological Diversification. (FRB Boston Working Paper 05/1.)

Levine, Ross and David Renelt (1992) A Sensitivity Analysis of Cross-Country Growth Regressions. American Economic Review 82:4, 942-963.

Li, H., L. Squire, and H. F. Zou (1998) Explaining International and Inter-temporal Variations in Income Inequality. Economic Journal 108, 26-43.

Lucas, R. E. J. (1988) On the Mechanics of Economic Development. Journal of Monetary Economics 22:1, 3-42.

Mobarak, Ahmed Mushfiq (2001) Why Democracies are More Stable: Models of Policy Choice and Political Conflict. University of Maryland. (Mimeographed).

Morduch, Jonathon (1995) Income Smoothing and Consumption Smoothing. Journal of Economic Perspectives 9:3, 103-114.

Nooruddin, Irfan (2003) Credible Constraints: Political Institutions and Growth Rate. Doctoral Dissertation, Department of Political Science, University of Michigan.

Obstfeld, M. (1994) Risk Taking, Global Diversification, and Growth. American Economic Review 84:5, 1310-1329.

Pallage, Stephane and Michel Robe (2003) On the Welfare Cost of Economic Fluctuations in Developing Countries. (forthcoming) International Economic Review.

Perotti, R. (1996) Growth, Income Distribution and Democracy: What the Data Say. Journal of Economic Growth 1,149-187.

Pritchett, Lant (2000) Understanding Patterns of Economic Growth: Searching for Hills among Plateaus, Mountains and Plains. World Bank Economic Review 14:2, 221-50.

Quinn, Dennis and John Woolley (2001) Democracy and National Economic Performance: The Preference for Stability. American Journal of Political Science 45:3,634-557.

Ramey, Garey and Valerie Ramey (1995) Cross-Country Evidence on the Link Between Volatility and Growth. American Economic Review 85:5, 1138-1151.

Rodrik, Dani (1999) Where Did All the Growth Go? External Shocks, Social Conflict, and Growth Collapses. Journal of Economic Growth 4:4, 385-412.

Saint-Paul, G. (1992) Technological Choice, Financial Markets and Economic Development. European Economic Review 36, 763-781.

Scheinkman, J. A., and W. Xiong (2003) Overconfidence and Speculative Bubbles. Journal of Political Economy 111, 1183-1219.

Stockman, A. C. (1988) Sectoral and National Aggregate Disturbances to Industrial Output in Seven European Countries. Journal of Monetary Economics 21,387-409.

Wolfers, Justin (2003) ls Business Cycle Volatility Costly? Evidence from Surveys of Subjective Wellbeing. (NBER Working Paper No. 9619.)

Comments

This paper examines the nature of volatility of GDP along with the volatility of different economic sectors of the economy, and focuses on the question as to what extent the volatility in the GDP growth rate is associated with the volatility of various sectors of the economy. The paper concludes that in a country like Pakistan, no long-run relationship exists between the volatility of GDP growth rate and the volatility of sectoral growth rates. However the relationship exists for the short run only.

As the title of the paper suggests, the authors were expected on the one hand to explain the microeconomic and macroeconomic determinants of sectoral volatility of GDP and on the other, to establish the linkages between sectoral volatility, economic development and governance. The contents of the paper, however, clearly show the failure of the authors to clearly diagnose the basic determinants of sectoral volatility and to establish the transmission mechanisms between sectoral volatility, development and governance. In fact, the discussion on these two critical aspects is almost non-existent in the paper.

The literature review in the paper refers to some important studies on the subject, however, the paper does not explain their findings and conclusions. The authors could develop a tabulated matrix to show the readers when each of these studies was conducted, what was the sample of the country data and what were the results. This would have significantly improved the contents of the paper.

The authors point out that this paper would answer a number of questions about the volatility of GDP growth rate etc. and then go on to pen down a few specific questions. However the answers to many of these questions are missing in the paper, which include the following:

(a) What are the main implications of the volatility parameters for Pakistan's policy problems'?

(b) What are the main implications of the volatility parameters for the achievement of stable growth rate of real G DP?

(c) How is the volatility in different sectors of Pakistan's economy correlated with each other?

(d) What is the nature of volatility of different economic sectors?

The regression results presented in the paper need far greater explanation than is provided in the paper. Simply stating that the impact variables are insignificant is not sufficient analysis. The idea is not to show the regression results, but to explain them in the light of changes in the stock and flow of economic variables. This could make the discussion interesting as well as meaningful. At the same time, there is a need to incorporate other independent variables in the regression equations which can be proxies for development and governance. Simply regressing the volatility in the overall GDP with the volatility in sectoral GDP is an oversimplified exercise and does not provide any useful policy conclusions.

In the sub-section on Impulse Response Function (IRF), each figure (Figure 5 - 9) in the paper requires exclusive explanation. The results from the IRF suggest that the sectoral volatility has no significant impact on the overall performance of the economy. One needs a decomposition analysis to validate such a result. There are various decomposition techniques in the literature which the authors could utilise for improving their results.

The literature on the theory and usage of Unit-Root Test and Angel-Granger Approach is well-known by now. The rudimentary explanation of these tests should be placed the in an appendix. There is no need for these details to be placed as a sub-section in the main body of a paper explaining policy implications of sectoral volatility.

In the concluding part of the paper, a reference has been made about restructuring the fiscal federalism in Pakistan. This is an unwarranted addition as the fiscal policy and demand management policies have not been incorporated in the econometric tests conducted in the paper. The reader is unable to see how sectoral volatility of the economy can be linked with fiscal federalism.

Aqdas Ali Kazmi

Planning Commission, Islamabad.

Toseef Azid is Fulbright Fellow at Social and Behavioural Sciences Division, El Camino College, Torrance, California, USA. Naeem Khaliq is a postgraduate student at Markfield Institute of Higher Education, Loughborough University, UK. Muhammad Jamil is Staff Economist at the Pakistan Institute of Development Economics, lslamabad.

Authors' Note: The views expressed are those of the authors and do no necessarily represent those of the institutions at which they study or work.

(1) In this most of Views regarding the patron of economic volatility in Pakistan are based on author's personal observation and experience of rescarch work.

(2) This is the only data source that provides quarterly data on GDP of Pakistan from 1971 to 2003 however State Bank of Pakistan also starts reporting quarterly figures on GDP since 1998 but that is not considered in the present study.

Appendix A Data of the Volatility Variables Obtained from
the ARCH-GARCH Process

 Volatility in Growth Rate (VOLT GR) of

Year VAG WIN VSER

1971-72-Q1
1971-72-Q2 3.9916
1971-72-Q3 -1.2284
1971-72-Q4 1.2416
1972-73-Q1 0.0916
1972-73-Q2 -0.5484
1972-73-Q3 -0.9184
1972-73-Q4 0.2716
1973-74-Q1 -1.0384
1973-74-Q2 6.9516
1973-74-Q3 -0.8484
1973-74-Q4 -0.0884
1974-75-Q1 -2.4784
1974-75-Q2 -0.0060 -2.7784 0.0966
1974-75-Q3 -0.0022 -1.0184 0.1487
1974-75-Q4 -0.0249 0.2716 0.0179
1975-76-Q1 -0.0058 -1.2684 0.0313
1975-76-Q2 0.0007 5.4616 -0.0642
1975-76-Q3 0.0480 -0.8184 -0.0311
1975-76-Q4 0.0587 -1.3084 -0.1132
1976-77-Q1 -0.0452 9.2316 0.0251
1976-77-Q2 -0.0461 0.0116 -0.0326
1976-77-Q3 0.0875 -1.3184 -0.0461
1976-77-Q4 0.0172 -1.3784 -0.0450
1977-78-Q1 0.0345 41.4016 0.0158
1977-78-Q2 0.0118 2.1616 0.0282
1977-78-Q3 0.0098 -1.1284 0.0421
1977-78-Q4 -0.0277 1.0616 0.0700
1978-79-QI -0.0105 -0.9884 -0.0119
1978-79-Q2 0.0188 1.1616 -0.0135
1978-79-Q3 0.0116 -0.7384 0.0194
1978-79-Q4 0.0409 -0.1884 0.0283
1979-80-Q1 0.0068 -1.2684 -0.0313
1979-80-Q2 0.0670 6.5516 0.0148
1979-80-Q3 -0.0585 -0.7484 0.0125
1979-80-Q4 -0.0275 -0.2884 0.0219
1980-81-Q1 -0.0375 -1.2784 -0.0181
1980-81-Q2 0.0480 6.5416 0.0061
1980-81-Q3 0.0303 -0.8484 0.0027
1980-81-Q4 0.0049 0.0416 0.0150
1981-82-Q1 -0.0093 -1.8884 -0.0479
1981-82-Q2 0.0419 -6.3684 0.0982
1981-82-Q3 0.0136 -0.9084 0.0276
1981-82-Q4 -0.0518 -0.9584 -0.0164
1982-83-Q1 0.0262 0.3916 0.0835
1982-83-Q2 0.0171 -0.5584 -0.0446
1982-83-Q3 -0.0244 0.2616 0.0272
1982-83-Q4 -0.0511 -0.9484 0.0146
1983-84-Q1 -0.0355 0.2216 0.0154
1983-84-Q2 -0.0926 0.0716 0.0320
1983-84-Q3 0.0340 -0.4784 0.0339
1983-84-Q4 -0.0083 -1.3784 -0.0520
1984-85-Q1 0.1004 -7.4484 -0.0544
1984-85-Q2 0.0495 -5.2884 0.0375
1984-85-Q3 -0.0856 -0.0884 0.0574
1984-85-Q4 0.0034 -1.2884 -0.0178
1985-86-QI 0.0725 1.6416 0.0338
1985-86-Q2 0.0077 -0.6884 -0.0525
1985-86-Q3 -0.0412 0.8016 -0.0430
1985-86-Q4 0.1109 -1.0184 0.0144
1986-87-Q1 0.0656 -0.1784 -0.0212
1986-87-Q2 0.0230 -0.1484 0.0154
1986-87-Q3 -0.0463 -0.0884 -0.0171
1986-87-Q4 -0.0211 0.9384 -0.0150
1987-88-Q1 0.0267 -0.1684 0.0062
1987-88-Q2 0.0321 -0.3684 -0.0030
1987-88-Q3 -0.0249 0.5316 0.0144
1987-88-Q4 -0.1909 -1.1384 -0.0140
1988-89-Q1 0.0651 0.7016 -0.0310
1988-89-Q2 0.0686 -0.6184 -0.0126
1988-89-Q3 -0.0215 0.9316 0.0076
1988-89-Q4 0.1018 -1.3584 -0.0344
1989-90-Q1 -0.0589 5.2416 -0.0227
1989-90-Q2 -0.0267 -0.2284 0.0039
1989-90-Q3 0.0057 -0.1684 -0.0247
1989-90-Q4 -0.0026 -0.8784 0.0137
1990-91-Q1 0.0416 -0.0884 -0.0137
1990-91-Q2 0.0193 -0.4684 0.0150
1990-91-Q3 -0.0190 0.0316 -0.0071
1990-91-Q4 0.0544 -0.8784 -0.0005
1991-92-Q1 0.0432 -0.3584 0.0093
1991-92-Q2 0.0510 -0.0684 0.0199
1991-92-Q3 -0.0274 -0.2684 -0.0012
1991-92-Q4 -0.0022 -0.6184 0.0088
1992-93-Q1 -0.0681 0.3416 0.0222
1992-93-Q2 -0.1037 -1.0484 -0.0254
1992-93-Q3 0.0238 0.2316 -0.0097
1992-93-Q4 -0.0049 -0.6884 -0.0015
1993-94-Q1 0.0280 -0.2984 -0.0144
1993-94-Q2 -0.0547 -0.2584 -0.0002
1993-94-Q3 0.0140 0.0516 0.0107
1993-94-Q4 -0.0481 -0.6084 -0.0009
1994-95-Q1 -0.0109 -0.3284 -0.0258
1994-95-Q2 -0.0372 -0.5584 0.0001
1994-95-Q3 0.0440 0.1716 0.0147
1994-95-Q4 0.0230 0.9184 0.0086
1995-96-Q1 0.0929 -0.0484 0.0002
1995-96-Q2 0.0418 -0.5084 0.0123
1995-96-Q3 0.0306 0.3016 0.0141
1995-96-Q4 0.0028 -0.9984 -0.0041
1996-97-Q1 -0.0011 -0.4484 -0.0222
1996-97-Q2 0.0022 0.2116 0.0157
1996-97-Q3 0.0153 0.1216 -0.0153
1996-97-Q4 -0.0087 -0.8984 -0.0299
1997-98-Q1 0.0080 -0.7384 -0.0427
1997-98-Q2 -0.0198 0.3216 -0.0082
1997-98-Q3 0.0428 -0.1284 -0.0285
1997-98-Q4 0.0193 -1.1384 -0.0345
1998-99-Q1 -0.0667 -0.1584 -0.0225
1998-99-Q2 -0.0233 2.6016 0.0544
1998-99-Q3 0.0098 -1.4184 -0.0805
1998-99-Q4 -0.0442 51.0916 0.0548
1999-00-Q1 -0.0296 -1.1584 0.0008
1999-00-Q2 0.0532 1.6116 -0.0221
1999-00-Q3 -0.0486 -0.8084 0.0319
1999-00-Q4 0.0652 0.6016 0.0083
2000-Ol-Q1 -0.0548 -1.1584 0.0098
2000-O1-Q2 -0.0531 3.3116 0.0006
2000-01-Q3 -0.0336 -1.2784 -0.0098
2000-01-Q4 -0.0732 3.9216 0.0216
2001-02-Q1 -0.0272 -0.8584 0.0137
2001-02-Q2 0.0326 0.9116 -0.0047
2001-02-Q3 -0.0502 -1.0684 -0.0098
2001-02-Q4 -0.0687 0.6116 0.0065
2002-03-Q1 -0.0081 -0.8484 0.0139
2002-03-Q2 -0.0049 0.1616 -0.0050
2002-03-Q3 0.0122 -0.7184 0.0203
2002-03-Q4 0.0084 -0.0684 -0.0079

 Volatility in Growth Rate (VOLT GR) of

Year VIN VWH Y

1971-72-Q1
1971-72-Q2
1971-72-Q3
1971-72-Q4
1972-73-Q1
1972-73-Q2 0.0171 -0.0010
1972-73-Q3 0.0742 0.0100
1972-73-Q4 0.0200 0.0003
1973-74-Q1 -0.0014 0.0158
1973-74-Q2 -0.0293 0.0252
1973-74-Q3 0.0276 0.0651
1973-74-Q4 0.0219 0.0342
1974-75-Q1 0.0199 -0.0504
1974-75-Q2 -0.0218 -0.0128 0.0215
1974-75-Q3 -0.0888 -0.0556 0.0335
1974-75-Q4 -0.0356 -0.0352 -0.0034
1975-76-Q1 0.0553 -0.0244 0.0255
1975-76-Q2 -0.0541 -0.0729 -0.0401
1975-76-Q3 0.0216 -0.0124 0.0401
1975-76-Q4 -0.0306 -0.0009 0.0021
1976-77-Q1 -0.0682 -0.1356 0.0004
1976-77-Q2 -0.0135 -0.0398 -0.0265
1976-77-Q3 -0.0004 0.0043 -0.0024
1976-77-Q4 0.0070 -0.0291 0.0033
1977-78-Q1 0.0017 0.0348 0.0443
1977-78-Q2 0.0499 0.0418 -0.0010
1977-78-Q3 0.0367 -0.0056 0.0119
1977-78-Q4 0.0041 -0.0184 0.0171
1978-79-QI 0.0015 -0.0332 -0.0221
1978-79-Q2 -0.0196 0.0071 -0.0034
1978-79-Q3 0.0083 0.0114 -0.0075
1978-79-Q4 0.0420 0.0316 0.0302
1979-80-Q1 0.0817 0.0365 0.0165
1979-80-Q2 -0.0147 -0.0025 0.0241
1979-80-Q3 0.0192 -0.0284 0.0084
1979-80-Q4 0.0112 0.0189 0.0181
1980-81-Q1 0.0098 -0.0104 -0.0151
1980-81-Q2 0.0418 0.0213 0.0029
1980-81-Q3 0.0358 0.0101 0.0053
1980-81-Q4 -0.0126 0.0182 -0.0111
1981-82-Q1 0.0443 -0.0004 -0.0377
1981-82-Q2 0.0241 0.0511 0.0325
1981-82-Q3 -0.0038 0.0112 -0.0005
1981-82-Q4 0.0732 0.0321 -0.0261
1982-83-Q1 -0.0265 0.0228 0.0063
1982-83-Q2 -0.1112 0.0201 -0.0504
1982-83-Q3 -0.0276 -0.0057 0.0072
1982-83-Q4 0.0211 0.0204 0.0058
1983-84-Q1 -0.0111 -0.0330 -0.0145
1983-84-Q2 0.0064 -0.0547 -0.0356
1983-84-Q3 0.0275 0.0691 0.0072
1983-84-Q4 -0.0139 -0.0154 -0.0344
1984-85-Q1 0.0314 0.0775 0.0116
1984-85-Q2 0.0340 0.0748 0.0272
1984-85-Q3 0.0315 -0.0253 0.0178
1984-85-Q4 -0.0929 -0.0208 -0.0285
1985-86-QI 0.0341 0.0113 0.0129
1985-86-Q2 0.0470 0.0284 0.0061
1985-86-Q3 -0.0071 -0.0363 -0.0088
1985-86-Q4 0.0089 0.0500 0.0068
1986-87-Q1 0.0167 -0.0061 0.0086
1986-87-Q2 -0.0070 0.0225 0.0298
1986-87-Q3 0.0526 0.0080 -0.0303
1986-87-Q4 0.0116 -0.0385 -0.0316
1987-88-Q1 -0.0048 0.0205 0.0238
1987-88-Q2 0.0254 0.0532 0.0089
1987-88-Q3 0.0640 0.0444 0.0165
1987-88-Q4 0.0034 -0.0027 -0.0161
1988-89-Q1 -0.0425 -0.0535 -0.0544
1988-89-Q2 -0.0393 0.0173 0.0105
1988-89-Q3 0.0001 0.0436 0.0194
1988-89-Q4 0.0321 0.0045 0.0015
1989-90-Q1 0.0078 0.0046 -0.0470
1989-90-Q2 0.0411 0.0052 -0.0078
1989-90-Q3 -0.0155 -0.0296 -0.0023
1989-90-Q4 -0.0575 -0.0077 -0.0099
1990-91-Q1 0.0090 -0.0042 -0.0146
1990-91-Q2 0.0096 0.0462 -0.0096
1990-91-Q3 0.0063 -0.0207 -0.0246
1990-91-Q4 0.0015 -0.0241 -0.0009
1991-92-Q1 0.0318 0.0774 0.0494
1991-92-Q2 -0.0104 -0.0029 0.0356
1991-92-Q3 0.0241 0.0100 -0.0051
1991-92-Q4 -0.0242 -0.0118 -0.0019
1992-93-Q1 -0.0257 -0.0403 -0.0162
1992-93-Q2 -0.0085 -0.0083 -0.0164
1992-93-Q3 0.0146 0.0021 0.0119
1992-93-Q4 -0.0052 0.0179 -0.0131
1993-94-Q1 -0.0248 0.0114 0.0103
1993-94-Q2 -0.0255 -0.0564 -0.0206
1993-94-Q3 0.0013 0.0004 -0.0049
1993-94-Q4 -0.0328 -0.0155 -0.0184
1994-95-Q1 0.0061 -0.0320 -0.0409
1994-95-Q2 -0.0581 0.0177 -0.0204
1994-95-Q3 0.0126 0.0019 0.0183
1994-95-Q4 0.0368 0.0710 0.0254
1995-96-Q1 -0.0006 -0.0248 -0.0078
1995-96-Q2 0.0201 0.0277 0.0150
1995-96-Q3 -0.0198 0.0237 -0.0131
1995-96-Q4 -0.0463 -0.0152 -0.0163
1996-97-Q1 -0.0162 -0.0230 -0.0230
1996-97-Q2 -0.0380 0.0465 -0.0059
1996-97-Q3 -0.0352 -0.0682 -0.0270
1996-97-Q4 -0.0819 -0.0810 -0.0502
1997-98-Q1 -0.0282 -0.0222 -0.0306
1997-98-Q2 0.0657 -0.0256 0.0071
1997-98-Q3 0.0477 -0.0236 0.0096
1997-98-Q4 -0.0478 -0.0055 -0.0182
1998-99-Q1 0.0263 -0.0100 -0.0184
1998-99-Q2 -0.0821 -0.0075 0.0237
1998-99-Q3 0.0210 -0.0166 0.0006
1998-99-Q4 0.0189 0.0274 0.0489
1999-00-Q1 -0.0248 0.0077 0.0036
1999-00-Q2 0.0005 -0.0542 0.0244
1999-00-Q3 -0.0885 0.0119 0.0047
1999-00-Q4 -0.0141 -0.0553 0.0310
2000-Ol-Q1 -0.0105 0.0700 0.0042
2000-O1-Q2 -0.0996 -0.1106 -0.0511
2000-01-Q3 0.0758 0.0744 -0.0108
2000-01-Q4 0.0124 0.0275 -0.0043
2001-02-Q1 0.0362 -0.0016 0.0022
2001-02-Q2 -0.0627 -0.0213 -0.0341
2001-02-Q3 0.0175 -0.1111 -0.0317
2001-02-Q4 -0.0173 0.0549 0.0012
2002-03-Q1 -0.0344 -0.0106 0.0115
2002-03-Q2 -0.0307 0.0603 0.0015
2002-03-Q3 0.0707 0.0384 0.0479
2002-03-Q4 -0.0216 -0.0206 0.0153

Table 1
Results of the Unit Root Tests

 ADF Test Statistics ADF Test Statistics

 Intercept Trend and Intercept

 First First
Variables Level Difference Level Difference

Y 2.868 -4.7467 * -2.447 -1.617 *
VAG 0.736 -5.7559 * -2.4458 -5.8665 *
VFIN -1.156 -9.9411 * -4.5852 * -9.9046 *
VIN 2.516 -5.9679 * -2.3260 -6.5399 *
VSER 5.148 -5.5177 * -0.8968 -7.4847 *
VWH 1.465 -5.1124 * -2.1241 -5.3080 *
VOL_Y -1.0489 -4.6774 * -3.4010 -4.6507 *
VOL_VAG -0.2862 -6.0048 * -3.4605 * -6.0116 *
VOL_VFIN -3.1330 * -6.0889 * -3.2373 -6.0790 *
VOL_VIN 0.0999 -6.0792 * -3.6675 * -6.2316 *
VOL_VSER -4.9313 * -6.1285 * -5.8330 * -6.1230 *
VOL_VWH -2.0712 -4.0367 * -2.1027 -4.0977 *
VOLL_Y -0.8132 -3.9776 * -4.4250 * -3.9310 *
VOLL_VAG -0.1096 -3.9744 * -3.3707 -4.0038 *
VOLL_VFIN -2.9116 * -5.5337 * -3.1435 -5.3472 *
VOLL_VIN 0.4487 -3.9141 * -3.7840 * -4.1450 *
VOLL_VSER -3.5416 * -4.3628 * -5.5585 * -4.3340 *
VOLL_VWH -2.3334 -2.8587 * -3.6414 * -3.8153 *

 ADF Test Statistics

 None

 First
Variables Level Difference

Y 6.3290 -1.6179
VAG 3.7258 -4.4730 *
VFIN 0.6380 -9.7911 *
VIN 7.3812 -2.8258 *
VSER 9.5974 -1.8768
VWH 6.0821 -2.7513 *
VOL_Y 0.9928 -4.4712 *
VOL_VAG 1.6703 -5.5789 *
VOL_VFIN -1.0052 -6.1098 *
VOL_VIN 1.2246 -5.7942 *
VOL_VSER -1.1751 -6.1443 *
VOL_VWH -0.5229 -4.0490 *
VOLL_Y -0.9980 -3.7275 *
VOLL_VAG 1.8916 -3.4342 *
VOLL_VFIN -0.8253 -5.4076 *
VOLL_VIN 1.6402 -3.3896 *
VOLL_VSER -0.6511 -4.3873 *
VOLL_VWH -0.4726 -2.8647 *

 ADF Test Statistics

 Trend and
 Intercept Intercept None

Variables Level Level Level

GRY -5.3590 * -5.8624 * -1.9516 *
GR_VAG -6.2150 * -6.1882 * -4.5707 *
GR WIN 1.7932 * -4.7726 * -4.4327 *
GR VIN -5.4886 * -5.9445 * -2.4740 *
GR_VSER -6.7988 * -7.0784 * -2.8983 *
GR_VWH -4.7395 * -5.0803 * -2.5758 *
VOLT_GRY -4.2038 * -4.2333 * -4.1701 *
VOLT_GR_VAG -4.8655 * -5.2690 * -4.8831 *
VOLT_GR_VEIN -4.7931 * -4.7726 * -4.6654 *
VOLT_GR_VIN -4.4653 * -4.8847 * -4.4841 *
VOLT_GR_VSER -5.2493 * -5.2392 * -5.2708 *
VOLT_GR_VWH -4.4865 * -4.4477 * -4.5006 *

Table 2
Regression Results Based on Different Volatility Variables

Dependent
Variable Constant VOLT_GR_VAG VOLT_GR_VEIN

VOLT_GR_Y -0.001 (0.497) 0.154 (0.001)
VOLT_GR_Y 0.002 (0.385) 0.001 (0.016)
VOLT_GR_Y -0.001 (1.580)
VOLT_GR_Y -0.002 (0.383)
VOLT_GR_Y -0.001 (0.619)
VOLT_GR_Y -0.002 (0.215) 0.154 (0.000) 0.001 (0.006)
VOLT_GR_Y 0.003 (0.162) 0.155 (0.000) 0.001 (0.006)
VOL_Y 716.586 (0.045) 1.418 (0.000)
VOL_Y 7585.577 (0.000) 2.000 (0.000)
VOL_Y 5803.860 (0.000)
VOL_Y 4070.161 (0.079)
VOL_Y 4110.778 (0.000)
VOL_Y 290.189 (0.187) 0.823 (0.000) -0.031 (0.784)
VOL_Y -282.798 (0.196) 0.877 (0.000) 0.825 (0.000)
VOL_Y -282.737 (0.193) 0.825 (0.000)
VOLL_Y 788.324 (0.007) 1.458 (0.000)
VOLL_Y 6495.665 (0.000) 2.561 (0.000)
VOLL_Y 4977.266 (0.000)
VOLL_Y 1114.281 (0.079)
VOLL_Y 2744.127 (0.000)
VOLL_Y -478.659 (0.007) 0.759 (0.000) -0.082 (0.415)
VOLL_Y -473.827 (0.008) 0.754 (0.000) -0.092 (0.366)
VOLL_Y -491.984 (0.000) 0.748 (0.000)

Dependent
Variable Constant VOLT_GR_VIN VOLT_GR_VSER

VOLT_GR_Y -0.001 (0.497)
VOLT_GR_Y 0.002 (0.385)
VOLT_GR_Y -0.001 (1.580) 2.286 (0.000)
VOLT_GR_Y -0.002 (0.383) 0.300 (0.000)
VOLT_GR_Y -0.001 (0.619)
VOLT_GR_Y -0.002 (0.215) 0.131 (0.006) 0.299 (0.000)
VOLT_GR_Y 0.003 (0.162) 0.137 (0.006) 0.304 (0.000)
VOL_Y 716.586 (0.045)
VOL_Y 7585.577 (0.000)
VOL_Y 5803.860 (0.000) 2.049 (0.000)
VOL_Y 4070.161 (0.079) 2.454 (0.000)
VOL_Y 4110.778 (0.000)
VOL_Y 290.189 (0.187) 0.455 (0.000) 1.040 (0.000)
VOL_Y -282.798 (0.196) 0.001 (0.995) 0.461 (0.000)
VOL_Y -282.737 (0.193) 0.461 (0.000) 1.008 (0.000)
VOLL_Y 788.324 (0.007)
VOLL_Y 6495.665 (0.000)
VOLL_Y 4977.266 (0.000) 2.286 (0.000)
VOLL_Y 1114.281 (0.079) 3.181 (0.000)
VOLL_Y 2744.127 (0.000)
VOLL_Y -478.659 (0.007) 0.416 (0.000) 1.190 (0.000)
VOLL_Y -473.827 (0.008) 0.418 (0.000) 1.203 (0.000)
VOLL_Y -491.984 (0.000) 0.420 (0.000) 1.451 (0.000)

Dependent
Variable Constant VOLT_GR_VWH DUMMY

VOLT_GR_Y -0.001 (0.497)
VOLT_GR_Y 0.002 (0.385)
VOLT_GR_Y -0.001 (1.580)
VOLT_GR_Y -0.002 (0.383)
VOLT_GR_Y -0.001 (0.619) 0.249 (0.000)
VOLT_GR_Y -0.002 (0.215) 0.120 (0.000)
VOLT_GR_Y 0.003 (0.162) 0.119 (0.011) 0.002 (0.506)
VOL_Y 716.586 (0.045)
VOL_Y 7585.577 (0.000)
VOL_Y 5803.860 (0.000)
VOL_Y 4070.161 (0.079)
VOL_Y 4110.778 (0.000) 3.703 (0.000)
VOL_Y 290.189 (0.187) 0.845 (0.000)
VOL_Y -282.798 (0.196) 1.008 (0.000) -192.99 (0.133)
VOL_Y -282.737 (0.193) -192.860 (0.125)
VOLL_Y 788.324 (0.007)
VOLL_Y 6495.665 (0.000)
VOLL_Y 4977.266 (0.000)
VOLL_Y 1114.281 (0.079)
VOLL_Y 2744.127 (0.000) 4.392 (0.000)
VOLL_Y -478.659 (0.007) 0.945 (0.000)
VOLL_Y -473.827 (0.008) 0.933 (0.000) 70.466 (0.387)
VOLL_Y -491.984 (0.000) 0.987 (0.000) 62.319 (0.441)

Dependent Adjusted
Variable Constant R-squared R-squared

VOLT_GR_Y -0.001 (0.497) 0.10089 0.09290
VOLT_GR_Y 0.002 (0.385) 0.05069 0.04228
VOLT_GR_Y -0.001 (1.580) 0.08862 0.08055
VOLT_GR_Y -0.002 (0.383) 0.21125 0.20427
VOLT_GR_Y -0.001 (0.619) 0.18274 0.17550
VOLT_GR_Y -0.002 (0.215) 0.05068 0.48415
VOLT_GR_Y 0.003 (0.162) 0.50880 0.48152
VOL_Y 716.586 (0.045) 0.87175 0.87071
VOL_Y 7585.577 (0.000) 0.18360 0.17696
VOL_Y 5803.860 (0.000) 0.76510 0.76319
VOL_Y 4070.161 (0.079) 0.47266 0.46837
VOL_Y 4110.778 (0.000) 0.44667 0.44217
VOL_Y 290.189 (0.187) 0.97582 0.97480
VOL_Y -282.798 (0.196) 0.97628 0.97507
VOL_Y -282.737 (0.193) 0.97628 0.97528
VOLL_Y 788.324 (0.007) 0.90629 0.90550
VOLL_Y 6495.665 (0.000) 0.21282 0.20620
VOLL_Y 4977.266 (0.000) 0.86371 0.86257
VOLL_Y 1114.281 (0.079) 0.64651 0.64354
VOLL_Y 2744.127 (0.000) 0.57882 0.57528
VOLL_Y -478.659 (0.007) 0.98840 0.98790
VOLL_Y -473.827 (0.008) 0.98848 0.98787
VOLL_Y -491.984 (0.000) 0.98839 0.98789

Dependent
Variable Constant F-statistic

VOLT_GR_Y -0.001 (0.497) 12.679 (0.0005)
VOLT_GR_Y 0.002 (0.385) 6.033 (0.0156)
VOLT_GR_Y -0.001 (1.580) 10.987 (0.0012)
VOLT_GR_Y -0.002 (0.383) 30.265 (0.0000)
VOLT_GR_Y -0.001 (0.619) 25 266 (0.0000)
VOLT_GR_Y -0.002 (0.215) 22.399 (0.0000)
VOLT_GR_Y 0.003 (0.162) 18.645 (0.0000)
VOL_Y 716.586 (0.045) 836.079 (0.0000)
VOL_Y 7585.577 (0.000) 27.662 (0.0000)
VOL_Y 5803.860 (0.000) 400.624 (0.0000)
VOL_Y 4070.161 (0.079) 110.245 (0.0000)
VOL_Y 4110.778 (0.000) 99.289 (0.0000)
VOL_Y 290.189 (0.187) 960.460 (0.0000)
VOL_Y -282.798 (0.196) 809.442 (0.0000)
VOL_Y -282.737 (0.193) 979.562 (0.0000)
VOLL_Y 788.324 (0.007) 1150.813 (0.0000)
VOLL_Y 6495.665 (0.000) 32.172 (0.0000)
VOLL_Y 4977.266 (0.000) 754.152 (0.0000)
VOLL_Y 1114.281 (0.079) 217.643 (0.0000)
VOLL_Y 2744.127 (0.000) 163.537 (0.0000)
VOLL_Y -478.659 (0.007) 1959.675 (0.0000)
VOLL_Y -473.827 (0.008) 1629.704 (0.0000)
VOLL_Y -491.984 (0.000) 1958.453 (0.0000)

Values within parenthesis are P-values.