SLIDE 1
Useful exergy is key in obtaining plausible APFs and in recognizing - - PowerPoint PPT Presentation
Useful exergy is key in obtaining plausible APFs and in recognizing - - PowerPoint PPT Presentation
1 Useful exergy is key in obtaining plausible APFs and in recognizing the role of energy in economic growth: Portugal 1960-2009 Science & Energy 2018 Joo Santos 4 th Science and Energy Seminar (joao.dos.santos@tecnico.ulisboa.pt) March 4
SLIDE 2
SLIDE 3
3
Neoclassical theory of economic growth
Solow, R. M. (1957). Technical change and the aggregate production function. The review of Economics and Statistics, 312-320. Swan, T. W. (1956). Economic growth and capital
- accumulation. Economic record, 32(2), 334-361.
Aggregate production function
- 0%
20% 40% 60% 80% 100% 1960 1970 1980 1990 2000 2010
≅ / ≅ /
Economic growth explained through accumulation
- f capital (), labor force growth (), and mostly
exogenous total factor productivity (). Energy resources are either downplayed or ignored altogether, and do not significantly contribute to economic growth.
- “Cost-share theorem”: factors of production are
paid according to their productive power.
Single-sector growth model
SLIDE 4
4
The role of energy in production: Ecological Economics
Aggregate production function
- Real-world economic processes cannot be fully understood without accounting for energy use, i.e.
energy is essential to production The economic system is embedded within a larger, environmental system, with interactions grounded on the laws of thermodynamics. The importance of energy to growth is higher than suggested by its cost share (! 10%).
Economy Environment
SLIDE 5
5
Aggregate Production Function critique
Felipe, J., & McCombie, J. S. (2005). How sound are the foundations of the aggregate production function?. Eastern Economic Journal, 31(3), 467-488. Felipe, J., & McCombie, J. S. (2013). The Aggregate Production Function and the Measurement of Technical Change: Not Even Wrong . Edward Elgar Publishing.
- Existence of an homogeneous degree one
APF linking output and inputs to production is often merely assumed;
- Conditions under which APF can be written
are stringent enough to doubt its existence;
- Aggregate measurement of capital inputs
implies adding up incomparable heterogeneous assets:
- Cambridge capital controversy;
“…the estimation of aggregate production functions is problematic, to say the least.” “…all those areas of neoclassical macroeconomics that use the aggregate production function (…) have no theoretical
- r empirical basis.”
SLIDE 6
6
Energy, exergy, and useful exergy
Electricity 0.08 kWh Lightbulb Power station Light 0.02 kWh Final Energy Primary Energy Waste Heat Heat 40% 60% 80% Final Exergy Primary Exergy 40%
20% 40% Useful Exergy
30% 30% Destroyed exergy
40% Destroyed exergy Heat Exergy
Coal (30 g) 0.27 kWh
Useful Energy
20% 30% Heat exergy
SLIDE 7
7
Exergy and useful exergy: Portugal 1856-2009
Serrenho, A. C., B. Warr, T. Sousa, R.U. Ayres, T. Domingos (2016). Structure and dynamics of useful work along the agriculture-industry-services transition: Portugal from 1856 to 2009. Structural Change and Economic Dynamics, 36, 1-21.
7 Final exergy / GDP Useful exergy / GDP Despite shifts in composition, useful exergy intensity is stable.
Final exergy / GDP (MJ/2010€) Useful exergy / GDP (MJ/2010€)
SLIDE 8
8
Exergy and useful exergy: Portugal 1960-2009
8
Useful Exergy / GDP Final Exergy / GDP Using a GDP-deflator instead of a Consumer Price Index, stability
- f UE intensity is clearer.
Final exergy / GDP (MJ/2010€) Useful exergy / GDP (MJ/2010€)
Serrenho, A. C., B. Warr, T. Sousa, R.U. Ayres, T. Domingos (2016). Structure and dynamics of useful work along the agriculture-industry-services transition: Portugal from 1856 to 2009. Structural Change and Economic Dynamics, 36, 1-21.
SLIDE 9
9
Cointegration: a drunk and her dog
1 2 3 4 5 6 1960 1970 1980 1990 2000 2010
Murray, M. P. (1994). A drunk and her dog: an illustration of cointegration and error
- correction. The American Statistician, 48(1),
37-39.
Drunk: %& ' %&() * +& Dog: & ' &() * ,& Each corrects his path so as not to stray too far from the other. Cointegration:
%& ' %&() * +& - . &() ' %&()
& ' &() * ,& - / %&() ' &()
SLIDE 10
10
Cointegration tests: Johansen* procedure and working variables
The order of integration can be tested resorting to unit root tests:
- Augmented Dickey-Fuller;
- Philips-Perron.
For time series with different order of integration, a set of 1 working variables can be defined to test for cointegration using the Johansen procedure:
* ln
- 2 * ln
- 3 * ln
- *Johansen, S. (1988). Statistical analysis of cointegration vectors.
Journal of economic dynamics and control, 12(2-3), 231-254.
The Johansen procedure to test for cointegration requires all time series to be 1. Economic output – 4 Capital inputs – 5 Labor inputs – Energy inputs – 6 Adoption of this set of working variables will impose constant returns to scale on APF formulations considered in our analysis. Likely to be 2
SLIDE 11
11
From cointegration to aggregate production functions
8)3 ' 892 ' : * 0 3 * ;< · 2 - :> * ? · @A · )(@A B, D The simplest model: 1) Cointegration between factor inputs and output; 2) Output elasticities must be positive and significant; 3) Granger causality between inputs and output; 4) Output elasticities E cost shares. Aggregate production function criteria Cointegrating relationship (vector):
- Output and capital (no energy);
- No linear time trend;
- At most one cointegration vector.
- * exp:> ·
- @A
?
Aggregate production function
SLIDE 12
12
From cointegration to aggregate production functions
8)3 ' 892 ' 8I ' : * 0 3 * ;< · 2 - ;J · - :> * ? · @A · @K · )(@A(@K B, D, L The simplest model: Cointegrating relationship (vector):
- Output, capital, and energy;
- No linear time trend;
- At most one cointegration vector.
- * exp:> ·
- @A
·
- @K
? 1) Cointegration between factor inputs and output; 2) Output elasticities must be positive and significant; 3) Granger causality between inputs and output; 4) Output elasticities E cost shares. Aggregate production function criteria
Aggregate production function
SLIDE 13
13
From cointegration to aggregate production functions
8)3 ' 89 ' : * 0 3 * ;J · - :> * ? · @K · )(@K The simplest model: Cointegrating relationship (vector):
- Output and energy (no capital);
- No linear time trend;
- At most one cointegration vector.
- * exp:> ·
- @K
? 1) Cointegration between factor inputs and output; 2) Output elasticities must be positive and significant; 3) Granger causality between inputs and output; 4) Output elasticities E cost shares. Aggregate production function criteria
Aggregate production function
- B, L
Aggregate production function
- B, L
SLIDE 14
14
From cointegration to aggregate production functions
8)3 ' 892 ' λN ' : * 0 3 * ;< · 2 - λ′N - :> * ? · expλ>N · @A · )(@A B, D The simplest model: Cointegrating relationship (vector):
- Output and capital (no energy);
- Linear time trend;
- At most one cointegration vector.
- * exp:> · expλ>N ·
- @A
? 1) Cointegration between factor inputs and output; 2) Output elasticities must be positive and significant; 3) Granger causality between inputs and output; 4) Output elasticities E cost shares. Aggregate production function criteria
Aggregate production function
SLIDE 15
15
From cointegration to aggregate production functions
P8)3 ' 892 ' :) * 0 8I2 ' 8Q ' :9 * 0 R * ?) · @A · )(@A * ?9 · S · )(S B, D, L The simplest model: Cointegrating relationship (vector):
- Output, capital, and energy;
- No linear time trend;
- At most two cointegration vectors.
- * exp:)
> ·
- @A
- * exp:9
> ·
- S
1) Cointegration between factor inputs and output; 2) Output elasticities must be positive and significant; 3) Granger causality between inputs and output; 4) Output elasticities E cost shares. Aggregate production function criteria
Aggregate production function
SLIDE 16
16
Cointegration: data for Portugal 1960-2009
Sources: AMECO / da Silva, E. G. & Lains, P. (2014). Capital formation and long-run growth: Evidence from Portuguese data, 1910-2011 (http://estructuraehistoria.unizar.es/personal/vpinilla/prog.htm) Sources: Penn World Tables 8.1 / Amaral, L. (2009). New Series for GDP per capita, per worker, and per worker-hour in Portugal, 1950-2007 (No. wp540). Universidade Nova de Lisboa, Faculdade de Economia.
2 4 6 1960 1970 1980 1990 2000 2010 2 4 6 1960 1970 1980 1990 2000 2010 3 6 9 12 1960 1970 1980 1990 2000 2010 1 2 3 1960 1970 1980 1990 2000 2010
Sources: Pinheiro, M. (1997). Séries Longas para a Economia Portuguesa, Pós-II Guerra Mundial, Vol. 1 – Séries Estatísticas, Lisbon, Banco de Portugal. Sources: Serrenho, A. C., Warr, B., Sousa, T., Ayres, R. U., & Domingos,
- T. (2016). Structure and dynamics of useful work along the agriculture-
industry-services transition: Portugal from 1856 to 2009. Structural Change and Economic Dynamics, 36, 1-21.
TUV 4 WLXYLW 5ZLX WYD 5ZD 4[\ ' ]^[WL] _ `a]^[WL] bc\ LaLd\ 6a
b
`WLe[ LfLd\ 6f
`
SLIDE 17
17
Cointegration: results and interpretation
,
g&hij klJmn,
,
gopqrios g&u
, ,
g&hij klJmn, v
,
gopqrios g&u
, v No cointegration. Model Cointegrating relationships found ∝ x.x9& · (x.xz · ).xz Implausible elasticity , {
|,
, }
~,
, {
|, v
, }
~, v
∝ x.x9& · (x.•€ · ).•€ Implausible elasticity No cointegration. No cointegration. No cointegration. No cointegration. No cointegration.
SLIDE 18
18
Cointegration: results and interpretation
,
g&hij klJmn, { |,
,
gopqrios g&u
, {
|,
,
g&hij klJmn, } ~,
,
gopqrios g&u
, }
~,
Implausible elasticity Model Cointegrating relationships found Elasticities not significant Implausible elasticity Implausible , elasticities ,
g&hij klJmn, { |, v
,
gopqrios g&u
, {
|, v
,
g&hij klJmn, } ~, v
,
gopqrios g&u
, }
~, v
∝ x.xx)& · x.)Q · x.zQ · x.99 ∝ (x.x9& · x.Qx · x.•• · (x.)• ∝ (x.xx)& · (x.)9 · x.€€ · x.9• ∝ (x.x)& · x.‚€ · (x.•) · x.•9 ∝ x.xx9& · (x.•€ · ).•9 · x.xz No cointegration. No cointegration. No cointegration. Implausible elasticity
SLIDE 19
19
Cointegration: results and interpretation
,
g&hij klJmn,
,
gopqrios g&u
, ,
g&hij klJmn, v
,
gopqrios g&u
, v No cointegration. Model Cointegrating relationships found ∝ (x.IQ · ).IQ Implausible elasticity ∝ x.zx · x.Qx ∝ x.€9 · x.)€ Implausible cost shares , {
|,
, }
~,
, {
|, v
, }
~, v
∝ x.Q• · x.•• Implausible cost shares ∝ x.•€ · x.99 ∝ x.€Q · x.)z ∝ x.•€ · x.99
SLIDE 20
20
Cointegration: results and interpretation
,
g&hij klJmn, { |,
,
gopqrios g&u
, {
|,
,
g&hij klJmn, } ~,
,
gopqrios g&u
, }
~,
Implausible elasticity Model Cointegrating relationships found Implausible cost shares Implausible , elasticities Implausible , , elasticities Implausible , elasticities ,
g&hij klJmn, { |, v
,
gopqrios g&u
, {
|, v
,
g&hij klJmn, } ~, v
,
gopqrios g&u
, }
~, v
ƒ ∝ x.zQ · x.Iz ∝ ).I)· (x.I) ƒ ∝ x.I• · x.zI ∝ 9.IQ· ().IQ ∝ x.•• · x.)z · x.9• ∝ x.€• · (x.z• · x.€9 ∝ (x.)• · x.€z · x.9‚ ∝ ).)€ · (9.x• · ).€‚ ∝ (x.•9 · ).•• · (x.xQ ƒ ∝ x.I) · x.z‚ ∝ I.)) · (9.)) Implausible elasticity
SLIDE 21
21
Cointegration: results and interpretation
* (‚.z•) · x.I) · x.z‚
0% 20% 40% 60% 80% 100% 1960 1970 1980 1990 2000 2010
For the “best” model: Capital-labor plausible APF: Estimated output elasticities compared with average historical cost shares for capital and labor.
- X. „%
- X. …„%
† ‡5 ≅ % † ‡ ≅ ˆ‰% For capital: For labor: Estimated output elasticities for capital and labor are remarkably similar to average historical cost shares for these factors of production. The neoclassical cost share theorem is compatible with this model.
SLIDE 22
- No economically plausible APFs are identified for models including a linear time trend;
- Unidirectional Granger causality running from energy inputs to economic output
supports the growth hypothesis;
- The only model satisfying all APF criteria is one with:
- a) no linear time trend;
- b) capital, labor, and energy inputs;
- c) quality-adjusted factors of production (capital services; human capital ajusted labor;
useful exergy);
- d) at most two cointegrating relationships among output and input variables.
- For this model:
- The first cointegrating relationship is normalized to output – capital-labor APF:
- The second cointegrating relationship – normalized to capital – expresses the real utilization
- f capital in production, as a function of labor and especially useful exergy:
22
Cointegration: conclusions
, * •.z‚ · I.)) · (9.))
SLIDE 23
- Expand the analysis to other countries, and/or groups of countries;
- Open up to alternative (more complex) APF formulations within the VAR cointegration
framework;
- Consider additional variables to include in the cointegration space;
- Test for normalizing and over-identifying restrictions to the cointegration space.
23
Cointegration: conclusions (cont’d)
- Adoption of a useful exergy metric to account for energy use provides important
insights to the relationships between energy use, macroeconomic factors of production, and economic output;
- A central role for useful exergy on economic production and growth is not incompatible
with the neoclassical assumptions of the cost-share theorem;
Cointegration: future work
Santos, J., Domingos, T., Sousa, T., St Aubyn, M. (accepted). Useful exergy is key in obtaining plausible aggregate production functions and recognizing the role of energy in economic growth: Portugal 1960-2009. Ecological Economics.
SLIDE 24
24
Acknowledgements
Financiado por fundos nacionais através da FCT/MCTES (PIDDAC) - projecto UID/EEA/50009/2013