Capital Goods Prices, Global "Capital Markets" and Growth: 1870-1950

William Collins

Department of Economics

Vanderbilt University

Nashville TN 37235

william.collins@vanderbilt.edu

 

Jeffrey G. Williamson

Department of Economics

216 Littauer Center

Harvard University

1805 Cambridge Street

Cambridge MA 02138-3001

Telephone: 617-495-2438

Fax: 617-495-7730

jwilliam@kuznets.fas.harvard.edu

The conventional wisdom now is that global financial markets were as well integrated in the 1890s as in the 1990s but that it took several post-war decades to regenerate the connections that existed before 1914. This view emerged from a variety of tests for world financial capital market integration ranging from the correlation of saving and investment aggregates (Feldstein and Horioka 1980; Taylor 1998) to the dispersion of securities prices and real interest rates (Neal 1985; Obstfeld and Taylor 1998). The ultimate importance of these connections has been justified in terms of the growth of nations and income convergence. Thus, when growth equations are applied to historical cross-sections from the Atlantic economy, measures of financial saving and access to foreign capital are often included (Prados et al. 1993; Obstfelt and Taylor 1998; O’Rourke and Williamson 1997, 1998: Chp. 11). Yet, financial capital markets cannot by themselves speak to the issue of growth without making explicit connections with capital goods.

A firm’s profits can be written: " = P*F(K, L) - (r + )*PK*K - wL, where P is the price of output, r is the real interest rate, is the depreciation rate, PK is the purchase price of a unit of capital, and w is the wage rate. Differentiating with respect to K yields a first order condition that implies that FK(K, L) = (PK/P)(r + ). Ordinarily, growth models assume P = PK = 1, but here we let them differ, as, of course, they did in fact. Let the output price be unity and ignore taxes for the moment. The user cost of capital can then be written u = PK(r+). Then du = [r+]dPK + Pkd[r+], and ultimately u* = PK* + (r+)*, where the "*" denotes growth rates. Changes in the user cost of capital have, therefore, two additive components, changes in the real interest rate (plus depreciation) and changes in the cost of capital goods, a fact that holds for changes over time as well as differences across countries. It is very surprising, therefore, that the relative price of capital goods has rarely been incorporated into studies of global capital market integration and long-term growth (except, as always, Kuznets 1961).

This could be an omission that matters. First, it is possible that the dispersion of the user cost of capital across the Atlantic economy was far wider and moved very differently than real interest rates alone suggest. If that were so, then the incentive to invest and accumulate would also have been far different than real interest rates suggest. Conventional tales about epochs of integrated and disintegrated capital markets may require some revision after evaluating trends in capital goods prices over time; so too, may assertions about the contribution of global financial capital markets to convergence and catch up in the late 19th century. Second, the relative price of capital goods (and machinery in particular) has been featured prominently in recent studies of post-war growth (De Long and Summers 1991; Jones 1994; Lee 1995). J. Bradford De Long and Lawrence Summers, for example, claim that "low equipment prices operate to promote growth by increasing the quantity of equipment investment" (1991: p. 474). For example, Argentina’s import-substitution policy in the 1940s and 1950s demonstrates how protection-induced high equipment prices can lead to low accumulation rates and slow growth (De Long 1992; Taylor 1992). Nonetheless, quantitative explorations of pre-1950 growth, between countries and through time, have been amazingly silent on the issue.

By combining capital goods and machinery price time series from national accounts with data from the Penn World Tables, this paper investigates the importance of capital goods prices from 1870 to 1950 for eleven countries, ten from the Atlantic economy, plus Japan. The paper begins by discussing the limitations in the data: Tables 1a and 2a are included in this summary to show the evolution in the price of capital goods relative to consumer goods, Pk, over time and relative to the United States. Tables documenting the relative price of equipment, Pm, are not included in this summary. The paper points out which countries had relatively cheap capital goods and when. It then explores the determinants of Pk and Pm. The paper moves on to show how conventional wisdom regarding "world capital market integration" must be revised with the rightful presence of PK. Finally, the paper offers some tentative assessments on the impact of relative capital goods prices on investment, accumulation and growth. Space constraints allow us to summarize only one of the main conclusions here. The price of capital goods relative to consumer goods was low in the United States compared with similar price ratios elsewhere in the Atlantic economy between 1870 and 1950. The US appears to have had that advantage in 1870, and it never lost it over the eighty years following. Relatively cheap capital goods made it easier for the US to increase capital-intensity and thus to defend its leadership position in the industrial peaking order. Without those cheap capital goods, the United States would have lost their industrial leadership sooner and faster, unless, of course, the US discovered others ways to offset the loss in that advantage.

Table 1a: Relative Capital Good Prices 1870-1950 (1900=100)

 

AUS

CAN

DEN

FIN

GER

ITA

JAP

NOR

SWE

UK

US

1870-74

79.08

84.98

88.16

87.70

105.59

100.30

-----

96.35

79.04

100.93

89.33

1875-79

76.90

87.80

84.48

84.06

85.84

103.53

141.99

100.24

82.86

97.06

90.41

1880-84

65.56

89.70

78.02

83.62

78.87

100.79

98.70

91.11

84.44

93.88

93.06

1885-89

61.28

90.22

87.85

94.21

84.97

97.82

118.35

92.61

86.18

90.90

93.20

1890-94

69.51

92.42

89.12

85.05

81.49

94.23

108.27

95.29

83.79

91.39

90.30

1895-99

80.34

97.16

97.77

97.81

86.47

87.02

107.33

96.42

90.64

93.02

95.44

1900-04

86.71

98.09

96.69

95.14

90.91

98.79

93.77

98.04

93.09

94.48

99.26

1905-09

82.80

93.32

98.49

93.93

89.14

98.30

85.87

99.62

91.94

90.72

99.91

1910-14

85.96

94.91

100.84

85.39

85.32

92.00

76.04

95.19

92.86

92.95

98.81

1915-19

107.92

-----

-----

83.78

-----

127.10

94.65

128.65

114.56

92.07

105.34

1920-24

109.10

108.74

108.48

97.63

-----

104.35

89.77

110.50

108.58

83.04

109.85

1925-29

106.91

98.81

98.65

106.16

91.41

94.28

75.89

82.62

100.21

73.35

106.19

1930-34

99.00

98.52

113.09

104.69

76.85

92.85

73.99

83.82

109.07

73.84

118.87

1935-39

107.85

103.82

110.60

118.47

88.62

87.17

79.78

91.26

113.51

78.26

120.01

1940-44

122.78

108.67

84.93

121.46

-----

51.68

85.82

-----

103.89

81.30

121.14

1945-49

142.92

111.05

106.77

131.10

-----

77.73

-----

119.97

104.62

97.30

119.41

1950

133.85

116.05

120.66

148.58

95.45

91.22

87.99

127.09

101.88

100.14

130.57

Notes: The figures show changes in the price of capital goods relative to consumption goods over time in each country. These prices are drawn from each country’s national accounts. Each series is scaled to equal 100 in 1900.

Table 2a: Relative Capital Goods Prices 1870-1950 (US = 100 in 1950)

 

AUS

CAN

DEN

FIN

GER

ITA

JAP

NOR

SWE

UK

US

1870-74

81.44

84.76

89.82

49.88

94.44

119.03

-----

78.54

94.56

125.34

68.42

1875-79

79.19

87.58

86.08

47.81

76.77

122.86

282.95

81.71

99.13

120.54

69.24

1880-84

67.51

89.48

79.50

47.56

70.54

119.61

196.69

74.27

101.02

116.59

71.27

1885-89

63.10

90.00

89.51

53.58

76.00

116.09

235.85

75.50

103.10

112.89

71.38

1890-94

71.58

92.19

90.80

48.37

72.88

111.82

215.77

77.68

100.24

113.49

69.16

1895-99

82.73

96.92

99.62

55.63

77.34

103.27

213.89

78.60

108.44

115.52

73.09

1900-04

89.30

97.85

98.51

54.11

81.32

117.23

186.87

79.92

111.36

117.34

76.02

1905-09

85.27

93.09

100.35

53.42

79.73

116.65

171.12

81.21

109.99

112.66

76.52

1910-04

88.53

94.68

102.74

48.56

76.31

109.17

151.53

77.60

111.09

115.44

75.68

1915-19

111.13

-----

-----

47.65

-----

150.83

188.62

104.87

137.05

114.34

80.67

1920-24

112.36

108.47

110.53

55.53

-----

123.83

178.90

90.08

129.89

103.13

84.13

1925-29

110.10

98.56

100.52

60.38

81.76

111.88

151.24

67.35

119.88

91.09

81.33

1930-34

101.95

98.28

115.23

59.54

68.74

110.18

147.45

68.33

130.49

91.70

91.04

1935-39

111.07

103.56

112.69

67.38

79.26

103.44

158.98

74.40

135.80

97.19

91.91

1940-44

126.44

108.40

86.54

69.08

-----

61.33

171.01

-----

124.28

100.96

92.77

1945-49

147.18

110.77

108.78

74.56

-----

92.24

-----

97.80

125.16

120.83

91.45

1950

137.84

115.76

122.94

84.50

85.37

108.25

175.34

103.60

121.88

124.37

100.00

Notes: The figures show the relative price of capital goods in each country compared to the relative price of capital goods in the United States in 1950. The movement of each series over time is determined by the price series of that country’s national accounts. Internationally comparable price data from the Penn World Tables are used to benchmark these national account series relative to the United States in 1950.