Introduction to NEG Steven Brakman, Harry Garretsen Regensburg, November 9, 2004

Transcription

1 Introduction to NEG Steven Brakman, Harry Garretsen Regensburg, November 9, 2004 Why study NEG? Why are traditional Regional Economics and International Trade theory models unsatisfactory Introduction in the core NEG model Some empirical Evidence

2 9 November Introduction. A first look at geography, trade, and development Introduction Clusters in the world economy Economic interaction Rapid change in population and production 2. Geography and economic theory Regional and urban economics Trade theory Conclusions: Stylized Facts

3 Clusters: global, continental, and national Nice life on the countryside

4 Clusters: global, continental, national Crowded life in the city (Dordrecht, the Netherlands, 2002)

5 Clusters: global, continental, national World population; total, rural, and urban (million) 6000 nonetheless: we increasingly live in (crowded) cities Total population Rural population Urban population

6 Figure. World Bank regional classification* global Regional classification World Countries by Column M EAP (2) ECA (27) LAC (33) MNA (4) SAS (8) SSA (48) HIGH (49) *EAP = East Asia and Pacific; ECA = Europe and Central Asia; LAC = Latin America and Caribbean; MNA = Middle East and North Africa; SAS = South Asia; SSA = Sub-Saharan Africa; HIGH = High-income countries.

7 igure.2 Life expectancy and regional shares of population, land, and income* a. Land area; 996 b. Population; 997 8% 4% 2% 8% % 22% 30% 8% 5% 25% 5% 8% 6% 8% EAP ECA High LAC MNA SAS SSA EAP ECA High LAC MNA SAS SSA

8 continental igure.4 Population density in Europe Population density people/sq km (996) 79 to 457 (6) 27 to 79 (5) 09 to 27 (5) 97 to 09 (5) 85 to 97 (3) 76 to 85 (6) 40 to 76 (5) 3 to 40 (6)

9 national Table. Ten largest urban agglomerations in India* Name Population Rank Ln(rank) Ln(population) Bombay 2, Calcutta, Delhi 8, Madras 5, Hyderabad 4, Bangalore 4, Ahmedabad 3, Pune 2, Kanpur 2, Lucknow,

10 Zipf's Law City size in India, 99 largest # people rank Bombay 2.5 million Calcutta million 2 ln(size) Delhi 8.4 million 3 etc. 20 Bombay Take ln(rank) and ln(size), make a plot 0 Calcutta Delhi ln( population R 2 = ) = ln( rank (528.4) ( 38.4) ln(rank) i i );

11 Distribution of economic activity at night

12 9 November, 2004 Introduction Chapter A first look at geography, trade, and development Introduction Clusters in the world economy Economic interaction Rapid change in population and production Chapter 2 Geography and economic theory Regional and urban economics Trade theory Conclusions

13 Inter-regional trade flows developing region The thickness of an arrow is proportional to the high income region size of the trade flow (percent of total interregional trade flows, rounded to nearest integer) LAC Latin America & Caribbean 5 SSA Sub-Sahara Africa 4 ECA East Europe & Central Asia NAm North America EUR West Europe SAS South Asia 4 EAP East Asia & Pacific AAs AustralAsia MNA Middle East & North Africa Charles van Marrewijk (2003) the slide show of this figure depicts inter-regional trade flows in decreasing order of magnitude. 44 of the 72 inter-regional trade flows are smaller than 0.5 per cent; these are not shown in the picture.

14 Table.3 Germany; 5 largest export markets, 998 Exports from Germany GDP Distance to Germany France 60.3, United States 5. 8,230 7,836 3 United Kingdom 46.3, Italy 40., Netherlands Belgium # Austria Switzerland Spain ,632 0 Poland Sweden ,259

15 Figure.7 German exports and distance, 998 monthly average Panel a. German exports 0 ln export y = -.472x R 2 = ln distance Panel b. German exports, income-adjusted ln adjusted exports y = x R 2 = ln distance

16 Gravity equation -4 Germany s largest export markets ln(export)-.033*ln(gdp) 5 Czech R. Austria 0 countryholland GDP 98 million US $ Distance to Germany Export from Germany France Switz United States Belgium 3 United Kingdom Italy Netherlands Belgium Austria Switzerland Spain Poland Japan 45-0 ln( export R 2 i = ) = ln(distance) ln( GDP) ln( distance ( 0.40) (34.86) i ( 2.77) i )

17 Figure.6 The Hard Disk Drive value chain* Component Fabrication Activities Subassembly Activities Final Assembly Activities Substrate Motors Media Head Fabrication Head- Gimbal Assembly Head-Stack Assembly Head-Disk Assembly Semiconductors PCB Stuffing and Other Electronics Assembly Final Assembly Testing *Simplified version of Gourevitch et al. (2 000), Figure.

18 Table.2 Hard Disk Drives; Indicators of nationality of production* Measure# US Japan S.E. Asia Other Asia Europe Other Nationality of firm Final assembly Employment Wages paid *Source: Gourevitch et al. (2000), Table 2 (data for 995). All numbers as percentage of world total. # Nationality of firm (% of unit output); location of final assembly; employment in value chain; and wages paid in value chain, respectively.

19 9 November 2004 Introduction Chapter A first look at geography, trade, and development Introduction Clusters in the world economy Economic interaction Rapid change in population and production Chapter 2 Geography and economic theory Regional and urban economics Trade theory Growth and development Conclusions

20 World population, UN projection (200): high 0000 estimate population (millions) low estimate realisation projection year

21 igure.8 Share of urban population in Europe* year Data source: Huriot and Thisse (2000, p. ix).

22 Introduction A first look at geography, trade, and development Introduction Clusters in the world economy Economic interaction Rapid change in population and production 2 Geography and economic theory Regional and urban economics Trade theory Conclusions

23 Regional economics structural modeling in neo-classical tradition; regions/cities Urban economics structural modeling in neo-classical tradition; cities, newer Economic geography Eclectic approach/heterodox economics; emphasis on empirics, outside influences (sociology, political science, regulation theory) Roots in German tradition: von Thuenen, Christaller, Weber, Loesch briefly discuss - von Thuenen; flat plane, city in the middle - central place system; flat plane, where cities? common feature: no endogenous explanation of size and location

24 Figure 2. The Von Thünen model a. Bid rent curves flowers A vegetables B grain flowers vegetables grain Distance from city center

25 b. Figure 2. continued flowers vegetables grain

26 Central place system villages: where to go for your bread? featureless plain with population

27 villages, small cities, large cities

28 Figure 2.4 Central places in a Dutch polder Lemmer Rutten Kuinre Urk Creil Espel Emmeloord Bant Luttelgeest Marknesse Tollebeek Kraggenburg Nagele Ens 0 km Kampen

29 Table 2. Population of locations in the Noord-Oost Polder* Location Start Planned population Population in 985 Emmeloord 946 0,000 8,976 Marknesse 946 2,000 2,94 Ens 948 2,000,68 Kraggenburg 948 2, Luttelgeest 950 2, Bant 95 2, Rutten 952 2, Creil 953 2, Nagele 954 2,000,04 Espel 956 2, Tollebeek 956 2,

30 6 July Introduction Chapter A first look at geography, trade, and development Introduction Clusters in the world economy Economic interaction Rapid change in population and production Chapter 2 Geography and economic theory Regional and urban economics Trade theory no role for geography (except endowments) Conclusions

31 Trade theory Neo-classical Emphasizes differences in relative factor intensity for the production of goods and services and differences in factor abundance for explaining the direction of trade flows - perfect competition - constant returns to scale New trade Emphasizes the importance of increasing returns to scale and the availability of many different varieties to explain (intra-industry) trade flows - imperfect competition - increasing returns to scale

32 The Neo-Classical Theory of Trade Here is a production plant for CDs (in China)

33 The Neo-Classical Theory of Trade This is the production rice (in Bangladesh) The production of CDs is relatively capital intensive and the production of rice is relatively labor intensive

34 Neo-classical economics International trade based on differences in endowments results Factor price equalization (FPE) Trade in goods (which equalizes final goods prices) leads to equalization of factor prices Rybczynski theorem (Ryb) An increase in the quantity of a factor of production at constant final goods prices leads to an increase in the production of the good using that factor intensively and a decreased production of the other good Stolper-Samuelson theorem (St-Sa) An increase in the price of a final good increases the reward to the factor used intensively in the production of that good and reduces the reward to the other factor Heckscher-Ohlin theorem (H-O) A country will export the final good which makes relatively intensive use of the relatively abundant factor of production

35 Problem: Intra Industry Trade A method to measure IIIT is the Grubel-Lloyd Index: The Grubel-Lloyd index measures the share of imports or exports (whichever is largest) that is covered by exports or imports of similar types of goods (or services); for cars in Holland in 996: E M = = = 0.22 = E + M Thus, 78% of the Dutch car trade is intra-industry trade GL-index = ; 00% Intra Industry Trade GL-index = 0; 0% IIT / 00% Inter Industry Trade (and then: Ricardo and/or HOS) Is Intra-Industry Trade relevant in our present-day economies?

36 Intra Industry Trade If we calculate the Grubel-Lloyd index not just for the Dutch car trade but for many sectors in for example Japan, Germany & USA, weigh them appropriately, we get the following results: Intra-industry trade (2-digits) 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0, Japan Germany USA

37 Intra Industry Trade; Variety/Intermediate Goods The basis for intra-industry trade is the love of variety characteristic in utility or production functions. - Welfare has increased dramatically over the past 00 years - We can consumer more of all the consumption goods than before - But an even bigger change in welfare is due to the fact we can consume more different and new goods than before

38 Intra Industry Trade; Variety/Intermediate Goods So how do we model this love of variety effect in order to analyse the outcomes? We can use an index of goods i =,..,n ; the index indicates which goods have been invented and introduced at a particular time. An increase in the index n therefore indicates an increase in the number of goods (varieties) available. U Utility (Dixit/Stiglitz, 977) Production (Ethier, 98) n = i = α c i α Increase of varieties (n) leads to an increase in welfare (utility) due to the love-ofvariety-effect. Q n = i= α q i 0 < α < 0 < α < α Increase of intermediate goods (n) leads to an increase in production due to positive production externalities.

39 Intra Industry Trade; Increasing Returns to Scale (IRS) If the production function requires a fixed labor cost f before production starts at variable labor cost m there are IRS: l = f + 29a i mx i 9 labor input total labor input average labor input output

40 Intra Industry Trade; Variety/Intermediate Goods If we now look at the Ethier-interpretation of IIT, what does that mean: -We do NOT look at utility and consumption of different (increasing) numbers of varieties of final goods, but - We look at the production of intermediate goods that occur through IRTS (CES function as before) Does international trade (IIT) follow from this? Yes, it does and it goes hand-in-hand with international production fragmentation. (not new, but strongly increasing) Because of IRTS in intermediate goods we see worldwide use of local comparative advantages to tap the advantages for international production (e.g. through Multinationals)

41 Intra Industry Trade; Variety/Intermediate Goods Intermediate A Intermediate B Intermediate 2 A. Intermediate N A Country A Producers country A Exports from B to A Exports from A to B Intermediate 2 B.. Intermediate N B Producers country B Country B Consumers country A Consumers country B

42 Intra Industry Trade; Variety/Intermediate Goods Example: Ford Fiesta Network engine components final assembly

43 igure 2.6 Share of world production in manufactures Share 0.6 S B A 0.4 S 2 Transport costs

44 9 November Introduction Chapter A first look at geography, trade, and development Introduction Clusters in the world economy Economic interaction Rapid change in population and production Chapter 2 Geography and economic theory Regional and urban economics Trade theory Conclusions

45 Conclusions Uneven distribution at all levels (global, continental, country) Economic activity more unevenly distributed than population Regularity in distribution (Zipf's Law) Regularity in interaction (Gravity equation) Objective: find common framework for understanding this Building blocks (990) economic geography/urban economics/regional science - existence and size of cities/regions is given international trade - no geography in core models economic growth and development - verbal analysis first nature second nature

46 3 The core NEG model An example Overview of general structure Demand Supply Equilibrium 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

47 Useful elements fromregional Economics Geometry, transport cost, but who is doing the optimizing? Physics of Location, gravity model identifies weights and pulleys of location (more on gravity model later) Cumulative causation, firms locate where markets are large, markets are large where firms locate, leds to multiple equilibria (not always optimal) Local externalities; technological, pecuniary (market size, thick labor market, etc) Economies of scale are important, what does

48 Two regions, production Man. Characterized by increasing returns to scale location is important. Each firm produces a single product Production Man., and workers mobile between regions Farmers immobile (always residual demand) Each firm sells 0 units-4 to workers, 6 to farmers(4 in region I, 2 in region II) Transportation costs;,-per unit Aim: minimize transportation costs A simple example as introduction

49 Table 3. Geography of sales Sales in North Sales in South Total Sales All firms in North = = 2 0 All firms in South = = % firms in North 75% firms in South + 4 = = 5 0 Table 3.2 Transport costs If location in North All firms in North = 2 (to farmers in South) All firms in South = 6 (to workers & farmers in South) 25% firms in North 75% firms in South = 5 (to workers & farmers in South If location in South = 8 (to workers & farmers in North) = 4 (to farmers in North) + 4 = 5 (to workers & farmers in North)

50 Illustrated in the example: cumulative causation multiple equilibria stable unstable non-optimality interaction between agglomeration and trade flows but: demand is given, no optimality decisions by individual agents, no determination of income, in short: no general equilbrium

51 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

52 Core model? What core model? Krugman (99); one factor of production, mobile labor Krugman and Venables (995); intermediate goods, no labor mobility Forslid and Ottaviano (200); two inputs, one input is mobile Main implications are all the same!

53 The structure of the model a Income (farm labor) Farm workers in (goods) Spending -δ m δ m Spending on food Consumers in Manufacturing workers in Income (labor) N manufacturing firms N varieties (elasticity ε) internal returns to scale monopolistic competition Spending on manufactures b f T g Mobility (λ i ) c Manufacturing workers in 2 Income (labor) Consumers in 2 N 2 manufacturing firms N 2 varieties (elasticity ε) internal returns to scale monopolistic competition e d Spending on manufactures Farm workers in 2 δ m (goods) Spending -δ m Spending on food Income (farm labor) Farms in Farms in 2 Direction of (goods and services flows) Direction of money flows

54 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

55 Demand manufactures - food max F, M Γ = F F δ δ M M δ δ + κ( Y s. t. Y = F + F IM ) IM δ δ δf M = κi δ ( δ )F M δ = κ IM = δf /( δ ) F = ( δ ) Y demand for manufacturing varieties max U = N i= N i= c c ρ i ρ (/ ρ ) / ρ = s. t. δy N i= p i c ( ρ ) / ρ Nc = N / ρ c = N (/ ρ ) i love of variety Nc claim on resources

56 + = Γ = = N i i i N i i c p Y c ) (/ δ κ ρ ρ N j for p c c j j N i i,..,, ) (/ = = = κ ρ ρ ρ ) /( ρ ε N j for c p p c or p p c c j j j j,.., = = = ε ε ρ ρ [ ] Y I p c or Y I c p p c p c p p p c p N j j N j j j N j j j δ δ ε ε ε ε ε ε ε ε, = = = = = = = = [ ] ) /(,, ε ε ε ε ρ ε δ = = N i i j j p I where Y I p c

57 Time for a Break?

58 igure 3.2 Dependence of demand for a variety of manufactures on price and ε* Differences in demand elasticity 200 demand for variety price of variety Demand given by ε c = 00 p ; ε varies (2, 4, and 6).

59 igure 3.3 Relationship between ρ and ε elasticity of substitution (epsilon) substitution parameter (rho)

60 igure 3.4 Deviation between assumed demand and reality a. N = 2 b. N = ,5 0,7 0,9,, ,5 0,7 0,9,,3 assumption reality assumption reality

61 Figure 3.4 continued c. N = 200 d. N = ,5 0,7 0,9,,3,8,6,4,2 0,8 0,6 0,4 0,2 0 0,5 0,7 0,9,,3 assumption reality assumption reality

62 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

63 Supply 0 ; ) ( < < = γ γ L F Food: labor as input, constant returns to scale, by choice of units ( ) ir x ir l β α + = Manufactures: different varieties, produced using only labor, internal increasing retursn to scale ( variety = firm); alpha, beta ) ( r r r r r x W x p β α π + = maximize profits ) / ( ) / ( ρ β β ε r r r r W p or W p = = optimal price setting: constant mark-up over marginal cost x x x W W x W x p x W x p r r r r r r r r r r r r = + = + = = + β ε α β α β ε ε β α β α ) ( ); ( ) ( 0; ) ( ( ) αε β ε α β α β α = + = + = ) ( x l ir exit or entry until zero profits

64 igure 3.5 Production function for a variety of manufactures* 9 labor input 6 3 total labor average labor output Fixed labor requirement α = 3, marginal labor requirement β =.

65 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

66 Multiple locations and equilibrium Total number of laborers; L (-γ) γ Laborers in the food sector (-γ)l Laborers in the manufacturing sector; γl φ φ 2 λ λ 2 Laborers in the food sector in region ; φ (-γ)l Laborers in the food sector in region 2; φ 2 (-γ)l Laborers in the manufacturing sector in region ; λ γl Laborers in the manufacturing sector in region 2; λ 2 γl Note: φ + φ 2 = Note: λ + λ 2 = Mobility Iceberg trade costs for manufactures; if T > are sent unit of distance only unit arrives (technology) Proportional to distance: if distance is 2, then T 2 have to be sent to ensure unit arrives (all sorts of distance impacts)

67 Table 3.3 CIF/FOB ratio s, (%) Country CIF/FOB ratio Country CIF/FOB ratio Australia 0.3 New Zealand.5 Austria 4. Norway 2.7 Canada 2.7 Philippines 7.6 Denmark 4.5 Portugal 0.3 France 4.2 Singapore 6. West-Germany 3.0 Spain 6.4 Greece 3.0 Sweden 3.5 Ireland 5.0 Switzerland.8 Italy 7. Thailand.0 Japan 9.0 United Kingdom 6.0 Netherlands 5.6 United States 4.9 Source: Radelet and Sachs (998).

68 Multiple locations and equilibrium (idem country 2) Short-run equilibrium; given the distribution of manufacturing labour Price index equation Income equation Wage equation (from demand = supply in I = λ W ε locally produced + λ T W ε ε 2 2 imported Y = λ W ) γ + φ ( γ manufacturing income food income /( ε ) [ ] ε ε ε / W ε manufactures sector = Y I + Y2T I 2

69 Multiple locations and equilibrium a. spreading b. agglomerate in region c. agglomerate in region region region 2 0 region region 2 0 region region 2 Three examples

70 Multiple locations and equilibrium Manufacturing labour force adjustment dλ = η ( w w); where w = λw + λ2w λ adj. wage average real change labor in speed difference 2 wage Table 4.2 When is a long-run equilibrium reached? Possibility Possibility 2 Possibility 3 If the real wage for manufacturing workers in region is the same as the real wage for manufacturing workers in region 2. All manufacturing workers are located in region (agglomeration in region ) All manufacturing workers are located in region 2 (agglomeration in region 2)

71 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

72 L L W Y i i i i ) ( γ φ γ λ + = ) /( ) ( ε ε ε ε λ αε γ ρ β = = R s rs s s r T W L I ε ε ε ε ρ α ε δ ρβ / / ) ( = = R r r rs r s I Y T W Table 4. Base-scenario parameter configuration, 2 regions δ = 0.4 γ = 0.4 L = ρ = 0.8 β = 0.8 φ = φ 2 = 0.5 T =.7 α = 0.08 σ = 0.000

73 Usefulness of computer simulations allows us to do things we cannot do analytically allows us to get a 'feel' for the model gives us suggestions we can try to prove analytically enables us to prove something does not hold: counter-example Computer simulations: how? Write down short version of model and determine endogenous variables (= # equations) specify the values of parameters determine a solution method; we use sequential iterations: o guess initial wage rates for and 2 o using this calculate income and price index for and 2 o use these to calculate possible new value for wage rates and 2 o repeat the above until a solution is found specify a 'stopping criterion' choose a programming language and write a program to do the above

74 Chapter 4 tool: computer simulations = short-run equilibria = long-run equilibria real wage region / real wage region 2 A B F C D E 0 share of manufacturing workforce in region

75 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change; T, rho, delta, gamma, alpha, beta, L, phi Normalization Intermezzo Sustain and break analysis Welfare Many locations

76 Chapter 4 tool: computer simulations relative real wage (w/w2). 0.9 T = 2. T =.7 T = share of manufacturing workers in region (lambda) T =.3 T =.7 T = 2. Krugman (99): "it.. assumes that w/w2 is a monotonic function.. or at least that it crosses one only once. In principle this need not be the case... I have not been able to rule this out analytically."

77 igure 4.2 The impact of transport costs Variations in transport costs T. w / w lambda

78 igure 4.3 The impact of some parameters panel a: variations in rho (substitution parameter) w / w lambda

79 Figure 4.3 continued Support site for the book: link to Forslid simulation panel b: variations in delta (share of income spent on manufactures) w / w lambda

80 Figure 4.3 continued panel c: variations in gamma (share of labor force in manufacturing).0 w /w lambda alpha, beta, L: nothing happens in the figure; suggestion for analysis

81 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Intermezzo Sustain and break analysis Welfare Many locations

82 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

83 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

84 Important: ask the right questions The sustainpoint is that configuration of structural parameters for which agglomeration of economic activity ceases to be a stable equilibrium. - Variables at agglomeration can be calculated. Determine implied real wage in periphery. Determine when this implied real wage is lower than in core. If so, sustainable. If not, not sustainable. - no-black-hole condition (rho > delta) The breakpoint is that configuration of structural parameters for which spreading of economic activity ceases to be a stable equilibrium. - Variables at spreading can be calculated. Suppose infinitesimal increase in lambda (= infinitesimal decrease in lambda 2). Determine all changes (advantage: changes in are exact opposite to changes in 2). Determine if real wage in increases. If so, unstable. If not, stable.

85 For the parameters δ and ρ: see Table 4.. Other parameters normalized, see Table 4.2. Note hat the agglomeration equilibrium is sustainable if f(t) <, see equation (4.5), while the preading equilibrium is stable if g(t) >, see equation (4.6). igure 4.4 Sustain-point and break-point I* 2 spreading stable,5 0,5 0 breakpoint sustain-point agglomeration sustainable.63.8,5 2 2,5 Transport costs T f(t) g(t)

86 Basin of attraction for agglomeration in region 2 Summary and hysteresis λ S 0.5 B 0 S 0 Transport costs T Sustain points Stable equilibria Break point Unstable equilibria Basin of attraction for spreading equilibrium Basin of attraction for agglomeration in region

87 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulatios and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

88 Figure 4.6 Welfare transport cost T lambda

89 Chapter 3 The core model An example Overview of general structure Demand Supply Equilibrium Chapter 4 Solutions and simulations Simulations and short-run equilibrium Structural change Normalization Intermezzo Sustain and break analysis Welfare Many locations

90 Figure 4.0 continued e. Initial and final distribution initial final preferred frequency analysis; frequency increases if (i) transport costs increase (ii) share of income spent on manufactures decreases, (iii) the elasticity of substitution increases.

91 Conclusions Combining various international economic theories with factor mobility provides a simple theory of location and agglomeration. Distinction short-run equilibrium (given distribution of the manufacturing labour force) and long-run equilibrium (endogenously determined by equality of real wages). Distinction stable equilibrium and unstable equilibrium. Using computer simulations: high transport costs lead to spreading of economic activity low transport costs lead to agglomeration of economic activity intermediate transport costs lead to multiple long-run equilibria Extensions of the basic model can explain empirical regularities, such as Zipf s Law and the Gravity Equation and economic viability of small industrial centers.

92 Empirical Evidence NEG Why testing is so difficult Example of estimating a NEG model An actual test: the Strategic Bombing of German cities during WW II

93 Empirical research NEG Three Examples - Agglomeration and factor prices (spatial wage structure) - Trade costs and agglomeration - Large shocks and stability of equilbria - [ Policy Implications ]

94 Puga 999 model - Crucial : interregional labor mobility? - Imperfectly elastic intraregional labor supply - Equilibr. wage equation (looks familiar!) W r [ ] ε ( ) { } µ ) ε ( ε ) I Y I T µ /( µ = Const r s s s rs

95 agglomeration a)interregional labor mobility: Tomahawk φ = ε b) No interregional labor mobility: rs T rs possibility of Bell Shaped Curve Ad b) Crux: spreading force independent from level of φ a+b: analytical expression break points

96 ½ O Φ S Φ B Φ, free - ness of trade degree of agglomeration

97 ½ O Φ B low Φ B high Φ, free-ness of trade degree of agglomeration

98 Levels, WLS 2 SLS Variable: ε 9,53 (6.9) Variable: γ 0.9 (22.)

99 Estimated φ meets φ- break -Given ε,γ and Drr: φ falls quickly when Drs -Break points: ) do not depend on distance specification ; 2) only function of model parametrs: ε ηµ δ -Given value of φ-break and our estimates: what are corresponding threshold distances??

100 Alessia Lazzarini Joëlle Tabak D rs /D rr φ B labmob φ B low φ B high

101 D rs φ B labmob φ B low φ B high

102 Interpretation of results -Compare threshold distances to real world distances: agglomeration forces are localized - What do other studies find? - Recall: Puga model is 2 region model!!

103 Conclusion - Agglomeration forces appear to be localized - How to proceed? - Better models. - Better testing. - An Example: The Strategic Bombing of German Cities during WW II

104 WHAT IS THE RELEVANCE OF NEG? AN EXPERIMENT

105

106 s = + + s α ( s s ) β i, 946+ t i,946 i,946 i,939 0 error i α = -, indicates a war effect, α = 0, indicates a random walk α β 0 Adj. R 2 West, t= (-4.03) 0.0 (0.35) 0.20 West, t= (-5.47) 0.05 (.87) 0.40 East, t= (0.88) (0.48) <0 East, t= (0.02) East+Wes t t= (-4.9) 0. (2.53) 0.07 (2.9) <0 0.3 CONCLUSION: For West, a temporary effect=>?