All Theses and Dissertations

Degree

Doctor of Philosophy in Economics

Faculty / School

School of Economics and Social Sciences (SESS)

Department

Department of Economics

Date of Award

Winter 2024

Advisor

Dr. Waliullah, Professor, Department of Economics, School of Economics and Social Sciences (SESS)

Second Advisor

Dr. Ehsan U. Choudhri, Chancellor’s Professor Emeritus and Distinguished Research Professor, Department of Economics, Carleton University - Canada

Committee Member 1

Dr. Waliullah, Supervisor

Committee Member 2

Dr. Abdul Jalil, Examiner I, National Defense University, Islamabad

Committee Member 3

Dr. Jameel Ahmed, Examiner II, State Bank of Pakistan

Project Type

Dissertation

Access Type

Restricted Access

Document Version

Final

Pages

xiv, 240

Keywords

Macro policy analysis, fiscal policy, monetary policy, DSGE model, VAR model, quarterly projections model, forecasting and policy analysis system

Abstract

Pakistan’s macroeconomic performance has been weak relative to its peer countries like India, China and Bangladesh. During the last two decades, indicators related to productivity, investments, exports, savings and tax mobilization have been on lower side in Pakistan. On the other hand, inflation, currency depreciation and debt accumulation have been on higher side.

Being the fifth most populous country with over 40 percent of poverty rate, Pakistan’s macroeconomic instability has enormous social costs, specially for the poor. We argue that improvement in macroeconomic policy formulation may bring about economic and financial stability in Pakistan.

Against this backdrop, objective of this dissertation is to contribute to macroeconomic policy formulation for Pakistan through development of models that are suitable for monetary and fiscal policy analysis. Economic models facilitate macroeconomic policy formulation in several ways. First, economic models build capacity in economic forecasting thereby assisting forward-looking policy analysis. Second, models with proper empirical properties support evidence-based policy analysis. Third, the use of sound models i.e. the models that are good in terms of theory, data and empirics, enhances transparency and credibility of policy making process. Finally, economic models improve communication of rationale of policy actions to general public, leading to increase in ownership of policy making process which ultimately translates into stability in economic expectations of general public.

Objective of this dissertation is to contribute to macroeconomic policy analysis in Pakistan by offering model-based policy inputs. To achieve this objective, we build various dynamic stochastic general equilibrium (DSGE) and vector autoregressive (VAR) models that are suitable for monetary and fiscal policy analysis in case of Pakistan. DSGE models are widely used for monetary and fiscal policy analysis by central banks, fiscal authorities, multilateral institutions and academia. We construct different versions of DSGE models and estimate them using Bayesian Maximum Likelihood (B-ML) estimation approach using quarterly data of macroeconomic variables for Pakistan economy. Estimation of VAR models vis a vis DSGE models is motivated by simplicity and sound empirical properties of these models. By the virtue of these properties, VAR models are generally considered the empirical benchmark models for macro policy analysis.

Popularity of DSGE models may be attributed to following desirable properties. First, these models offer comprehensive and holistic macroeconomic analysis owing to their general equilibrium structure. Second, their structural nature; rooted in microeconomics, makes sure that policy analysis offered by these models conforms to economic theory. This property facilitates communication of policy output to policy makers as well as general public. Third, advanced versions of these models possess comparable data fit and forecasting performance. Finally, statistical advantages of these models include identification of shocks, consistent estimation of mis-specified models and ability to overcome Lucas’ Critique.

In first essay of the dissertation, we construct and estimates an open economy DSGE model and apply the model for analysis of expenditure based fiscal policy analysis. The basic open economy framework utilized in the model is a simplified version of the model presented by Adolfson et al. (2007a). The model is an open economy extension of the closed economy DSGE framework offered by Christiano et al. (2005) and Smets and Wouters (2003). The model includes propagation mechanisms like sticky prices & wages, habit persistence in consumption behaviour, investment adjustment costs and variable capacity utilization. To assess the implications of financial frictions for monetary and fiscal policy shocks, we augment the model with financial accelerator mechanism offered by Bernanke et al. (1999). However, results of B-ML estimation show that inclusion of financial accelerator mechanism leads to deterioration of data fit in case of Pakistan.

To thoroughly evaluate implications of expenditure based fiscal policy, we model government consumption and government investment expenditures separately following Leeper et al. (2010). This framework allows government investments in public infrastructure to positively affect productivity of private sector, leading to realization of so-called “crowding-in” effects. Another important modelling feature of government investments is distinction between “authorized” vs. “implemented” government investment and modelling of investment implementation delay.

To utilize the model for fiscal policy analysis, we estimate present value multipliers of government consumption and investment over different forecast horizons. Our estimation results show that in the medium term, present value multipliers for government consumption and government investment expenditures are almost zero in case of Pakistan. Impulse response functions show that positive shocks to government consumption and investment lead to crowding out of private consumption, investment and exports, nullifying the positive impact of expansionary fiscal policy shocks.

Forecast evaluation exercise shows that the model is capable of producing forecasts of key macroeconomic variables that are comparable with Bayesian VAR models in the medium run.

In second essay, we estimate VAR models under different identification strategies, functional forms and specifications to estimate fiscal policy multipliers for Pakistan using quarterly data from 2003Q1 to 2024Q1. Our finally chosen specification of VAR model yields cumulative multipliers for total spending that are found to be negative; both in the short run and in the long run. Positive values of total spending multiplier are realized only during fifth to tenth quarters after introduction of spending shock. Long run value of cumulative spending multiplier becomes negative because initial gains in output are outweighed by subsequent negative effects of crowding out. On the other hand, net tax multiplier is negative in the short run but attains positive value in the second year and remains positive till four years after introduction of shock, implying positive effects of fiscal consolidation on real GDP. Disaggregated model results show that government consumption multipliers are larger than government investment multipliers. Government investment multipliers are negative in the long run. Based on these results, we show that use of expenditure based expansionary fiscal policy to boost output growth has no empirical support in case of Pakistan. Fiscal consolidation may support long run growth prospects in Pakistan.

In third essay, we present B-ML estimation of quarterly projections model (QPM) for Pakistan. The model, customized to improve data fit and forecasting performance over medium term, is actually rooted in a multi-sector DSGE model. The framework, initially introduced by Berg et al. (2006a) and Berg et al. (2006b) constitutes the core model employed by Forecasting and Policy Analysis Systems (FPAS) that are being utilized by many developing and developed countries for monetary policy analysis.

Pakistan has already been using a calibrated version of this model (Ahmad and Pasha, 2015). The third essay of this dissertation improves forecasting performance and relevance for policy analysis by offering B-ML estimation of the existing calibrated model. Estimation results based on quarterly data from 2001Q3 to 2023Q1 show substantial differences in values of estimated versus calibrated parameters related to aggregate demand, aggregate supply, monetary policy rule and exogenous shock processes. Comparison of pseudo out-of-sample forecasting performance for key macro variables shows that estimated model provides more precise forecasts in case of headline inflation, real GDP growth, interest rate and exchange rate over 8-quarters forecast horizon.

We use the estimated model for gap analysis and scenario analysis. Gap analysis shows that currently, Pakistan is passing through a recession with overshot exchange rate. In scenario analysis, we incorporate implications of political instability, climate risks, commodity prices and global financial conditions for next three years forecasts of domestic variables under baseline and alternate scenarios. Results of scenario analysis, which may also be used for scenario design in a macro stress testing exercise, show that simultaneous realization of assumed risk factors may lead to substantial deterioration of macroeconomic outlook. We explore different monetary and fiscal policy options to counter recent crisis. Our results show that under current circumstances, using an expansionary monetary policy may lead to substantial rise in inflation and macroeconomic volatility without offering sustainable gains in GDP growth.

Although presented as separate articles yet, the three essays of this dissertation are closely linked. Being related to different aspects of Pakistan’s macro-economy, like fiscal policy and monetary policy, strong links may easily be established among the findings of the three essays of the thesis. For instance, estimates of multipliers from VAR and DSGE models suggest that fiscal policy has only weak and short-lived effects on GDP growth. The main reasons behind this ineffectiveness of fiscal policy are scarcity of savings and consistent increase in fiscal deficit which have led to strong crowding out of private sector credit and investment. Against this backdrop, the third essay offers quite sensible policy prescription that the central bank should adopt tight monetary policy stance to incentivize private savings. An encompassing monetary policy scenario is necessary to ensure positive value for real interest rates, which in turn is critical to boost savings.

Download

Available for download on Saturday, March 20, 2027

The full text of this document is only accessible to authorized users.

Share

COinS