LIMITS TO GROWTH? Carl-Johan Dalgaard Department of Economics - PowerPoint PPT Presentation

LIMITS TO GROWTH? Carl-Johan Dalgaard Department of Economics University of Copenhagen

INTRODUCTION Classical writers in economics, like Thomas Malthus, saw land as an essential and fi xed factor of production. He also argued the size of population rises, if income goes up. As consequence of these two premisses he argued growth must come to a halt; population would over the long-run be kept at a level of subsistence Basic logic of the Malthusian “trap” . Suppose Y t = X β L 1 − β , where X is the fi xed supply. Observe t µ X ¶ β y t ≡ Y t = (1) L t L t for which reason L ↑⇒ y ↓ . 2

INTRODUCTION Moreover, Malthus argued that population size (fertility) was endoge- nous. If y ↑⇒ n ↑ . To capture this is a simple way, suppose n t = s n · y t , s n < 1 . (2) Remainig part (1 − s n ) is consumed. To equations (1) and (2) we add L t +1 = n t L t , n > 0 . (3) To analyse the model (equation 1-3) substitute for n and y into (3) L t +1 = s n y t L t = s n Y t = s n X β L 1 − β ≡ G ( L t ) , t which is the law of motion for population, aggregate output, and output per capita. [Insert phase diagram] 3

INTRODUCTION The steady state level of population L t +1 = L t = L ∗ = s 1 /β X. n Thus more land would sustain greater numbers of individuals. But, they would not be more “wealthy”: Ã ! β X y ∗ = = s − 1 n . s 1 /β X n Hence, “accumulation” of land would not permanently be able to im- prove living standards. The reason is (1) diminishing returns to labor input (conseqence of land entering into the production function), and (2) n increases with income 4

INTRODUCTION It was not long after Malthus completed his thesis that (2) ( n t = s n · y t ) started breaking down. Today, in rich places anyway, rising income does not lead to population growth. Rather it is other way around. BUT, if land indeed is important (i.e, present) in the production func- tion it may modify our results from the basic Solow model. Note that. Y t = AK α L β t X κ , α + β + κ ≡ 1 thus µ K t ¶ α µ X ¶ κ y t ≡ Y t = A L t L t L t As L rises the last term declines, pushing in the direction of lower standards of living. ISSUE 1 : Under what circumstances can growth be sustained? 5

ISSUE 1: LIMITS TO GROWTH? Consider a standard Solow model with technological change, augmented to include land. Slightly augemented replication argument ...: µ K t ¶ α µ X ¶ κ µ K t ¶ α 1 − α µ X ¶ κ β 1 − α y t = A β 1 − α ⇔ y t = A t L t L t Y t L t Otherwise the model is standard. That is, we have k t +1 = sy t + (1 − δ ) k t ⇒ k t +1 = sy t /k t + (1 − δ ) (1 + n ) k t (1 + n ) To solve the model in a simple way, de fi ne s/z t +(1 − δ ) k t +1 z t ≡ K t ⇒ z t +1 (1+ n ) k t = = y t +1 ³ A t +1 ´ β 1 − α ³ z t +1 ´ α 1 − α ³ L t ´ κ Y t z t y t 1 − α z t A t L t +1 6

LIMITS TO GROWTH? DYNAMICAL ANALYSIS After some rearrangements ⎛ ⎞ 1 − α z t +1 s/z t + (1 − δ ) ⎝ ⎠ = ≡ Φ ( z t ) β z t 1 − α − κ 1 − α (1 + n ) (1 + g ) 1 − α Which is the law of motion for capital intensity in the model. Observe that 1 − α − κ = β by constant returns to capital, labor and land. [ Insert phasediagram ] 7

LIMITS TO GROWTH? DYNAMICAL ANALYSIS β β + κ > (1 − δ ) a Hence, contingent on the condition [(1 + n ) (1 + g )] steady state exists, and it is stable. s z ∗ = > 0 . β β + κ − (1 − δ ) [(1 + n ) (1 + g )] What about growth in GDP per capita? µ y t +1 ¶ ∗ µ A t +1 ¶ β 1 − α µ z ∗ ¶ α 1 − α µ L t ¶ κ β = (1 + g ) 1 − α 1 − α = κ z ∗ y t A t L t +1 (1 + n ) 1 − α ³ y t +1 ´ ∗ Note: > 1 requires y t β κ 1 − α > (1 + n ) (1 + g ) 1 − α That is, only if the rate of technological change is su ffi ciently rapid is growth sustainable!(note: 1 − α = β + κ by CRTS) 8

LIMITS TO GROWTH? With land entering the production function we have reached the fol- lowing conclusion: Even with technological change growth in GDP per capita is not nessesarily sustainable, if the popu- lation expands expotentially One may view this as a “Neo-Malthusian” result: There are limits to growth. At some level perhaps uninteresting: it is obvious that popu- lation growth itself cannot go on inde fi nitely. Still, worth noting that “technological change” is not nessesarily enough to ensure growth, with limited resources and a rising population. Either y or n will have to “give in”. Comparative economic growth? 9

LIMITS TO GROWTH? “New” predictions (compared to standard Solow) : (1) The steady state growth rate of GDP per capita is negatively a ff ected by population growth. Intuition: familiar capital dillution mechanism “on steroids” " # 1 µ y t +1 ¶ ∗ (1 + g ) β β + κ = (1 + n ) κ y t (2) More land increases GDP per capita in the long-run " # 1 µ X ¶ κ µ X ¶ κ (1 + g ) β β β β + κ 1 − α α β + κ α y ∗ = A ( z ∗ ) ( z ∗ ) 1 − α 1 − α 1 − α = A 1 − α (1 + n ) κ t 0 L t L 0 ½ ¾ β with z ∗ = β + κ − (1 − δ ) s/ [(1 + n ) (1 + g )] . Both predictions are consistent with cross-country data (cf textbook). 10

NATURAL RESOURCES...A CURSE? We just saw that land is “good” for long-run living standards But there are other forms of natural resources which relate directly to production: Oil and mineral extraction in particular. Both are (for practical purposes) as nonrenewable natural resources. Hence, as the resource is used the stock of it declines. Oil, for instance, is used in production for its value as an energy input. What are the implications of admitting exhaustible natural ressources into the model? 11

NATURAL RESOURCES...A CURSE? The simplest version of the model has the following production function ³ ´ K α t L 1 − α Y t = min , A t E t t where E is energy. (textbook asumes Cobb-Douglas. (1) more com- plicated, (2) substitution of E for K with A given ... ultimately not meaningful from a thermodynamical perspective). Hence, we require µ K t ¶ α A t e t ≡ E t = k α t ≡ for all t. L t L t Now, suppose E t = s E R t , s E < 1 . where s E is the extraction rate. Finally, suppose R t +1 = (1 − s E ) R t , where R is the stock of the resource (i.e., oil). 12

NATURAL RESOURCES...A CURSE? Studying the dynamics y t s K k t + (1 − δ ) k t +1 = k t 1 + n R t +1 = (1 − s E ) R t de fi ne s K /z t +(1 − δ ) k t +1 z t ≡ K t ⇒ z t +1 (1+ n ) k t = = y t +1 A t +1 R t +1 /L t +1 Y t z t y t A t R t /L t where it has been used that k α t = A t e t = A t s E R t /L t at all points in time . z t +1 = s K /z t + (1 − δ ) (1 + g ) (1 − s E ) ≡ Φ ( z t ) z t 13

NATURAL RESOURCES...A CURSE? After some rearrangements, the law of motion reads z t +1 = s K /z t + (1 − δ ) (1 + g ) (1 − s E ) ≡ Φ ( z t ) z t Φ (0) = ∞ and Φ ( ∞ ) = (1 − δ ) / [(1 + g ) (1 − s E )] < 1 if δ > s E . With this condition, the phasediagram looks very much like the one for the model with land in the production function Steady state capital-output ratio: s K z ∗ = (1 + g ) (1 − s E ) − (1 − δ ) > 0 . 14

NATURAL RESOURCES...A CURSE? What about long-run growth? At any point in time k α t = A t e t We have shown, k/y = z ∗ . So rewriting the above t ) α = A t e t ⇒ y ∗ t = z ∗− 1 ( A t e t ) 1 /α ( z ∗ y ∗ µ y t +1 ¶ ∗ ∙ ¸ 1 /α µ (1 + g ) (1 − s E ) ¶ 1 /α (1 + g ) s E R t +1 /L t +1 = = y t s E R t /L t 1 + n Once again we have that technological change is not su ffi cient for eco- nomic growth to last. Without (energy saving) technological change, growth must come to a halt. Note : This is true even if population growth is absent ( n = 0) . 15

NATURAL RESOURCES...A CURSE? µ y t +1 ¶ ∗ µ (1 + g ) (1 − s E ) ¶ 1 /α = y t 1 + n The intuition is as with land, only the rate at which the resource be- comes dilluted is increasing in s E . This is why the resource extraction rate lowers long-run growth. The more “general” speci fi cation where substitution is possible softens this conclusion slightly; changes in s E and n does not map into changes in growth on a 1:1 basis (see textbook). The way in which we have introduced technology into the model makes clear that energy saving technological change are needed, if we do not rely on “substitution”. Is this process possible forever? 16