a new keynesian model for the analysis of energy shocks in pakistan by dario debowicz, ifpri
DESCRIPTION
Presentations made at the PSSP First Annual Conference - December 13, 14, 2012 - Planning Commission, Islamabad, PakistanTRANSCRIPT
A New Keynesian model for the analysis of energy shocks in Pakistan
Dario Debowicz
(International Food Policy Research Institute) and
Alejandro Quijada (Inter-American Development Bank)
December, 2012
Motivation: energy crisis in Pakistan
We realize the suffering that load shedding causes our people. We are painfully aware of the darkness it spreads, how children study by candlelight, and how the wheels of industry often stop. President Asif Ali Zardari's Speech at the Joint Session of Parliament Islamabad, April 5, 2010 The scale of the present crisis is formidable and requires persistent structural and pricing reforms in the sector, increased implementation of loss-reduction programs, and expanded investments. The government recognizes the need for substantial external assistance to help overcome the energy crisis. Pakistan Yearly Energy Book 2010
Some stylized facts
- The world price of crude oil and LNG increases more than 300% from 2000
- Energy is perceived as a bottleneck constraining growth and employment opportunities in Pakistan (Friends of Democratic Pakistan Energy book 2010).
- The energy constrain leads to a low rate of utilization of the capital stock (State Bank of Pakistan 2012).
- The energy sector of Pakistan is financially unsustainable today. Notified electricity tariffs are below the cost recovery level as per determined tariffs, so the government therefore subsidizes tariffs by providing tariff differential subsidies in the budget (Friends of Democratic Pakistan Energy book 2010).
World energy prices (US dollars per physical unit)
Source: World Bank Commodity Price Data (βPink sheetβ) BBL: oil barrel. MMBTU: Million Metric British Thermal Units
β’ World crude and gas prices grow by more than 300% from the start of 2000.
0
20
40
60
80
100
120
140
16020
00M
0120
00M
0720
01M
0120
01M
0720
02M
0120
02M
0720
03M
0120
03M
0720
04M
0120
04M
0720
05M
0120
05M
0720
06M
0120
06M
0720
07M
0120
07M
0720
08M
0120
08M
0720
09M
0120
09M
0720
10M
0120
10M
0720
11M
0120
11M
0720
12M
01
Crude oil, Dubai ($/bbl) Liquefied natural gas, Japan ($/mmbtu)
Net barter terms of trade, export and import value index (2000=100)
Source: World Development Indicators (World Bank)
β’ The countryβs terms of trade deteriorate during the last decade as the import prices (led by energy prices) increase significantly more than export prices. β’ Overall, the countryβs terms of trade fall almost 40% from 2000 to 2010
050
100150200250300350400450
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Export value index (2000 = 100)
Import value index (2000 = 100)
Net barter terms of trade index (2000 = 100)
Energy related subsidies in Pakistan (millions of rupees)
Source: PSSP elaboration based on 2011-12 βFederal budget in briefβ
β’ Outstanding energy-related subsidies are implemented, exceeding 372 billions of Pakistan rupees in 2010-11, and explaining more than half of the 680-billion fiscal deficit.
2006-07 2007-08 2008-09 2009-10 2010-11 2011-12WAPDA Budget 41,934 52,893 74,612 62,903 84,000 122,700
Revised 42,464 113,658 92,840 147,005 295,827 - KESC Budget 13,938 19,596 13,800 3,800 20,447 28,588
Revised 17,699 19,596 18,800 32,521 64,447 - Oil Refineries/OMCs Budget 10,000 15,000 140,000 15,000 10,807 7,921
Revised 25,000 175,000 70,000 11,224 10,807 Fertilizer Manufacturers Budget 980 10,360 12,860 210 185 162
Revised 978 6,360 21,268 439 985 - Total Budget 66,852 97,849 241,272 81,913 115,439 159,371
Revised 86,141 314,614 202,908 191,189 372,066 -
Fiscal balance (share of GDP)
(8)
(7)
(6)
(5)
(4)
(3)
(2)
(1)
-FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09
Source: Handbook of Statistics of Pakistan
β’ Recurrent large fiscal deficit, which exceeds 7% of GDP in FY 2008
Money supply (yearly growth)
Source: Broad Money Supply, Handbook of Statistics of Pakistan
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
β’ Through fiscal deficit monetization, money supply accelerates without interruption during last decade, reaching a 6% rate of growth in 2010
Load shedding in Pakistan
YearNational sales (GWH)
National load shedding (GWh)
Total National Demand (GWh)
Load Shedding (%)
2003 52,661 52,661 0.0%2004 57,467 520 57,986 0.9%2005 61,247 265 61,512 0.4%2006 67,608 1,208 68,815 1.8%2007 71,947 2,040 73,982 2.8%2008 72,518 12,578 85,096 14.8%2009 69,668 18,222 87,890 20.7%2010 73,595 21,821 95,238 22.9%
β’ Increasing load shedding from 2005, with alarming level of 22.9% in 2010
Real Gross Domestic Product growth (at market prices of 1999-00)
Source: Handbook of Statistics of Pakistan
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
FY01 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 P
β’ GDP growth was relatively low during most of the last decade considering population and labor supply growth due to contemporaneous population dividend, with energy bottleneck presumably having an important role.
Inflation (GDP deflator % changes)
Source: Handbook of Statistics of Pakistan
0.0
5.0
10.0
15.0
20.0
25.0
FY 00 FY 01 FY 02 FY 03 FY 04 FY 05 FY 06 FY 07 FY 08 FY 09 FY 10
β’ With expansionary macroeconomic (fiscal and monetary) policies and the supply bottleneck generated by energy, domestic inflation accelerated significantly in last years
Real effective exchange rate
0
20
40
60
80
100
120
14019
9019
9119
9219
9319
9419
9519
9619
9719
9819
9920
0020
0120
0220
0320
0420
0520
0620
0720
0820
0920
10
Source: International Financial Statistics (IFS) - IMF
β’ As domestic inflation systematically exceeded the sum of nominal devaluations and foreign inflation, the countryβs real effective exchange rate deteriorated significantly during the last twenty years
Current account balance
Source: Handbook of Statistics of Pakistan β’ Low competitiveness leads Pakistan to face a recurrent trade deficit during the last 30 years. β’ To a large extent, remittances helped in reducing the associated current account deficit. β’ Trade balance deteriorates in parallel to the energy crisis
Central questions
β’ How are the increases in world energy prices affecting growth, employment and inflation in Pakistan? Which are the major transmission channels?
β’ How are the energy-related domestic subsidies affecting domestic growth, employment and inflation? Which are the major transmission channels?
β’ How would an increase in the supply of energy impact growth, employment and inflation in the country? Which are the major transmission channels?
Building a New Keynesian model
β’ To answer these questions, we need a model that, capturing the channels by which energy-related shocks and macroeconomic policies affect the domestic economy, is able to shows us how the domestic economy is affected over time by changes in world energy prices, domestic energy subsidies, and domestic energy production.
β’ We frame the questions into the New Keynesian paradigm as:
β it captures the short-run interaction between macroeconomic policy, inflation, and the business cycle, accounting for agents inter-temporal planning and expectations.
β it allows us to look into how the economy responds over time to a set of relevant impulses through the analysis of impulse-response functions.
Main characteristics of New Keynesian model we are building
The model has production, consumption, international trade and financial markets. Production The economy produces output using labor and physical capital. In turn, in order to utilize the physical capital stock, producers demand energy. As a result, the use of energy affects the use of the capital stock, in turn affecting the productivity of labor and, ultimately, the output that the economy can produce. Reflecting the main components of the energy mix in Pakistan, energy is produced using oil and gas. Imperfect (monopolistic) competition characterizes the domestic factor and commodity markets. Consumption Households maximize inter-temporal utility, which depends on consumption, real balanced held for transaction purposes, and labor effort. Increases in the real interest rate lead households to postpone consumption over time.
Main characteristics of New Keynesian model we are building
International trade Exports (imports) are determined by global demand (domestic absorption) and the real exchange rate. Financial markets The model has explicit markets for money, domestic bonds and foreign bonds. Households allocate their portfolio among a set of financial and physical assets (money, domestic and foreign bonds and capital stock), and rent capital services to firms. As usual in New Keynesian models, the real interest rate is determined by monetary policy via the use of a Taylor rule, by which the Central Bank lifts the interest rate when output grows above steady-state growth, and/or when domestic inflation exceeds the steady-state one. The domestic interest rate, the foreign interest rate and the exchange rate are inter-linked through the uncovered interest parity condition, such that in equilibrium the expected return of domestic and foreign financial assets coincide.
β’ Calibrate the model based on macroeconomic data and information specific to the
energy sector (Handbook of Statistics of Pakistan, Pakistan Energy Book 2011), using among other data subsidy data that the Pakistan Bureau of Statistics provided us with.
β’ Implement simulations regarding world energy prices, domestic energy subsidies and
productivity in the domestic energy sector.
β’ Generate and analyze impulse-response functions that show us the time-path by which the economy returns to its steady-state equilibrium path after relevant shocks.
β’ Get feedback and fine-tune the analysis
Next steps
An, S and Kang, H. (2011), βOil Shocks in a DSGE Model for the Korean Economyβ, Chapter in NBER book Commodity Prices and Markets, East Asia Seminar on Economics, Volume 20 (2011), Takatoshi Ito and Andrew K. Rose, editors (p. 295 - 321). Adam, C. OβConnell, S and Buffie, E. (2008), βAid volatility, monetary policy rules and the capital account in African economiesβ, WEF Working Papers 0037, ESRC World Economy and Finance Research Programme, Birkbeck, University of London. Almeida, V. (2009), βBayesian Estimation of a DSGE Model for the Portuguese Economyβ, Bank of Portugal Working Papers Series No. 14/2009. Calvo, G. (1983), βStaggered Prices in a Utility-Maximizing Frameworkβ, Journal of Monetary Economics, Elsevier, vol. 12(3), pages 383-398, September. Choudhri, E.U. and H. Malik (2012), βMonetary Policy in Pakistan: A Dynamic Stochastic General Equilibrium Analysisβ, mimeo. Economist (2011). Lights out. Pakistan energy shortage.
Bibliography
Bibliography Friends_of_Democratic_Pakistan and Energy_Sector_Task_Force (2010). Integrated Energy Sector Recovery Report & Plan. Gali, J (2008), βMonetary Policy, Inflation and the Business Cycle, An Introduction to the New Keynesian Frameworkβ, Princeton University Press. IMF (2010). Pakistan: Poverty Reduction Strategy Paper. Medina, JP and Soto, C. (2005), βOil shocks and Monetary Policy in an Estimated DSGE Model For a Small Open Economyβ, Central Bank of Chile, Working Paper 353. Peiris, S and Saxegaard, M. (2007), βAn Estimated DSGE Model for Monetary Policy Analysis in Low-Income Countriesβ, IMF Working Paper 07/282. Washington DC: International Monetary Fund. World Bank. (2011a), βResponding to Global Food Price Volatility and Its Impact on Food Securityβ, Document for the Development Committee, DC2011-002, April 4, 2011, Washington DC: World Bank. World Bank. (2011b), βFood Price Watchβ, Poverty reduction and Equity Group. August 2011. Washington DC: World Bank.
Ξ0 β Ξ²tUt jβt=0 = Ξ0 β Ξ²t log ct j β ΞΆhctβ1 j + s
Β΅mt jpt
Β΅β Οt
lt j 1+Οl1+Οl
βt=0 (1)
ππ‘ π + ππ‘ π1+ππ‘
+ etππ‘β π
1+itβ Ξππ‘ππ‘
β
πππ‘ππ‘
= ππ‘β1 π + ππ‘β1 π + etππ‘β1β π + π€π‘ π ππ‘ π +
ππ‘ π ππ‘π π + πΉπ‘ π + πππ‘ π β ππ‘ππ‘ π β ππ‘ππ‘ π β ππ‘ππ‘πππ§π§,π‘(π) (2) ποΏ½ππ‘ β βΞΆπ½ποΏ½ππ‘+1 = π½ 1 + ππ‘
ππ‘ππ‘+1
ποΏ½ππ‘+1 β βΞΆπ½ποΏ½ππ‘+2 (3)
1 + itβ Ξ ππ‘ππ‘β
πππ‘ππ‘
ππ‘+1ππ‘
= 1 + ππ‘ (4)
ποΏ½ππ‘ββΞΆπ½ποΏ½ππ‘+1
ππ‘β π½ ποΏ½ππ‘+1ββΞΆπ½ποΏ½ππ‘+2
ππ‘+1= π
ππ‘
ππ‘ππ‘
πβ1 (5)
Model equations*
* The full description of the model is available in DSGE_Pakistan_Model.docx.
ππ‘ π = πΏ1π πππ‘ π
πβ1π + 1 β πΏ
1π π§π‘ π
πβ1π
ππβ1
(6)
π§π‘ π = πΎ1π πππ‘ π
πβ1π + 1 β πΎ
1π πβπ‘ π
πβ1π
ππβ1
(7) ππππ‘πππ‘ π + ππ§π‘π§π‘ π (8) πππ‘πππ‘ π + πβπ‘πβπ‘ π (9)
π§π‘ π = ππ§π‘πππ‘
βπ 1βπΏπΏπππ‘ π (10)
πβπ‘ π = πβπ‘πππ‘
βπ 1βπΎπΎπππ‘ π (11)
Model equations
ππ‘ = πΏππππ‘1βπ + 1 β πΏ ππ§π‘1βπ1
1βπ (12)
ππ§π‘ = πΎπππ‘1βπ + 1 β πΎ πβπ‘
1βπ1
1βπ (13)
ππ‘ = β« lt jΞ΅lβ1Ξ΅l dj
10
Ξ΅lΞ΅lβ1
(14)
ππ‘ π = π€π‘ ππ€π‘
βΞ΅lππ‘ (15)
π€π‘ = β« wt j 1βΞ΅ldj10
11βΞ΅l (16)
π€π‘+π π = Ξw,t
i wt j (17)
Ξw,ti = β ΞΆ 1 + Οt+jβ1
ΞΎl 1 + ΟοΏ½t+j1βΞΎli
j=1 (18)
Model equations
Ξt β π½ππ πππ‘+π π ποΏ½ππ‘+πΞ΅l
Ξ΅lβ1+ wt+i j
pt+iποΏ½ππ‘+π β βΞΆπ½ποΏ½ππ‘+1+πβ
π=0 = 0 (19)
ππ‘π (π) = π§π§π‘ π ππ‘β1(π) (20) ποΏ½ππ‘ββΞΆπ½ποΏ½ππ‘+1ποΏ½ππ‘+1ββΞΆπ½ποΏ½ππ‘+2
=
π½ ππ‘+1π§π§π‘+1ππ‘+1
+ 1 β πΏπ β ππ§π§1+ππ§π§
π§π§π‘+11+ππ§π§ β π§π§π π 1+ππ§π§ β πππππ§π§π‘+1ππ β
ππ‘+1π ππ§π§π§π§π‘+1
ππ
ππ
(21) ππ‘ππ‘β ππ§π§π§π§π‘ππ§π§ β ππ‘ππππ§π§π‘ππβ1 = 0 (22)
ππ‘ π = ππ‘ π + 1 β πΏπ ππ‘β1 π β Ξzu,t (23) Ξzu,t = ππ§π§
1+ππ§π§π§π§π‘ π 1+ππ§π§ β π§π§π π 1+ππ§π§ ππ‘β1(π) (24)
Model equations
ππ§π§,π‘(π)ππ‘β1(π)
= πππππ§π§π‘ π ππ (25)
πππ‘ = πππ‘ 1 β πΌπ1ππππππ‘
ππβ1ππ + πΌπ
1ππππππ‘
ππβ1ππ
ππππβ1
(26) πππ π‘ππππ‘
= πΌπ1βπΌπ
ππ,π‘ππππ,π‘
ππ (27)
π¦βπ‘ π = πβπ‘ 1 βΞ±1πππ‘ π
πβ1π + πΌ
1πππ‘π (π)
πβ1π
ππβ1
(28) ππ‘π(π )ππ‘ π
= πΌ1βπΌ
π€π‘ππ‘
π (29)
πππ‘ π = πππ‘ = 1πβπ‘
1 β πΌ π€π‘1βπ + πΌπππ‘1βπ1
1βπ (30)
Model equations
πΉπ‘ π = πβπ‘ π π¦βπ‘ π β π¦βπ‘ π πππ‘ (31)
π¦βπ‘ = β« yht sΞ΅hβ1Ξ΅h ds
10
Ξ΅hΞ΅hβ1
(32)
π¦βπ‘ π = πβπ‘ π πβ
βΞ΅hπ¦βπ‘ (33)
πβπ‘ = β« pht s 1βΞ΅hds10
11βΞ΅h (33β)
πβπ‘+π π = Ξh,t
i pht s (34) Ξh,ti = β 1 + Οht+iβ1 ΞΎh 1 + ΟοΏ½t+i 1βΞΎhi
j=1 (35) Ξt β π½ππ πππ§π‘+ππ¦βπ‘+π π 1 β Ξ΅h Ξh,t
i πβπ‘ π + Ξ΅hπππ‘+πβπ=0 = 0 (36)
Model equations
πβπ‘β = πΎβ πβπ‘β
πππ‘β
βπβ
ππ‘β (37) ππππ‘ = ππ‘πππ‘β
ππ‘ (38) ππ‘ = ππ‘
ππ‘β1β 1 (39)
πππ‘ = 1 + ππ ππ‘πππ‘β (40) πππ‘ = 1 + ππ ππ‘πππ‘β (41) πππππ‘ = 1 + ππππ ππ‘πππππ‘β (42) πβπ‘β = πβπ‘
ππ‘ (43)
ππ‘π
= ππ‘β1π
πππ¦π‘
π¦π‘β1ΞΆοΏ½
Ξ¨y 1βππΟtποΏ½
Ξ¨Ο 1βππ (44)
ππ‘ = πΎπ
1ππ πππ‘
ππβ1ππ + 1 β πΎπ
1ππ πβπ‘
ππβ1ππ
ππππβ1
(45)
πβπ‘ = πβπ‘πππ‘
βππ 1βπΎππΎπ
πππ‘ (46)
Model equations
ππ‘ππ‘ = π‘π‘ + ππ‘ β 1 + ππ‘ ππ‘β1 + ππ‘ β ππ‘β1 (47) π‘ππ‘ = ππ
1+ππππ‘ππ‘ (48) π‘ππ‘ = ππ
1+πππππ‘ππ‘ (49) π‘ππππ‘ = ππ
1+πππππππ‘ππππ‘ (50)
π‘πππ‘ = πππ
1+πππππππ‘πππ‘ (51) π‘πππ‘ = πππ
1+πππππππ‘πππ‘ (52)
π‘π‘ = π‘ππ‘ + π‘ππ‘ + π‘ππππ‘ + π‘πππ‘ + π‘πππ‘ + ππππ‘ππ‘ (53) π¦βπ‘ = πβπ‘ + πβπ‘ + πβπ‘β (54) πππ‘ = πππ‘ + ππ§π§,π‘ (55) ππ‘ = πβπ‘ + πππ‘ + πππ‘ (56) ππ‘ = πβπ‘ + πππ‘ (57) ππππ‘πππ‘ = πππ‘πππ‘ + πππ‘πππ‘ + πππ‘ππ‘ + πππππ‘ππππ‘ (58) ππππ‘ = π ππππ‘ + π ππππ‘ + π ππππππ‘ (59) ππ¦π‘π¦π‘ = ππ‘ππ‘ + ππ‘ππ‘ + ππ‘ππ‘ + πβπ‘πβπ‘β β ππππ‘πππ‘ (60)
etππ‘β
1+itβ Ξππ‘ππ‘
β
πππ‘ππ‘
= etππ‘β1β + πβπ‘πβπ‘β β ππππ‘πππ‘ (61)
Model equations