Baumol, Engel, and Beyond: Accounting for a century of structural transformation in Japan, 1885-1985

by Kyoji Fukao (Hitotsubashi University) and Saumik Paul (Newcastle University and IZA)

The full article from this blog post was published on The Economic History Review, and it is now available on Early View at this link

Bank of Japan, silver convertible yen. Available on Wiki Commons

Over the past two centuries, many industrialized countries have experienced dramatic changes in the sectoral composition of output and employment. The pattern of structural transformation, depicted for most of the developed countries, entails a steady fall in the primary sector, a steady increase in the tertiary sector, and a hump shape in the secondary sector. In the literature, the process of structural transformation is explained through two broad channels: the income effect, driven by the generalization of Engel’s law, and the substitution effect, following the differences in the rate of productivity across sectors, also known as “Baumol’s cost disease effect”.

At the same time, an input-output (I-O) model provides a comprehensive way to study the process of structural transformation. The input-output analysis accounts for intermediate input production by a sector, as many sectors predominantly produce intermediate inputs, and their outputs rarely enter directly into consumer preferences. Moreover, an input-output analysis relies on observed data and a national income identity to handle imports and exports. The input-output analysis has considerable advantages in the context of Japanese structural transformation first from agriculture to manufactured final consumption goods, and then to services, alongside transformations in Japanese exports and imports that have radically changed over time.

We examine the drivers of the long-run structural transformation in Japan over a period of 100 years, from 1885 to 1985. During this period, the value-added share of the primary sector dropped from 60 per cent  to less than 1 per cent, whereas that of the tertiary sector rose from 27 to nearly 60 per cent in Japan (Figure 1). We apply the Chenery, Shishido, and Watanabe framework to examine changes in the composition of sectoral output shares. Chenery, Shishido, and Watanabe used an inter-industry model to explain deviations from proportional growth in output in each sector and decomposed the deviation in sectoral output into two factors: the demand side effect, a combination of the Engel and Baumol effects (discussed above), and  the supply side effect, a change in the technique of production. However, the current input-output framework is unable to uniquely separate the demand side effect into forces labelled under the Engel and Baumol effects.

Figure 1. Structural transformation in Japan, 1874-2008. Source: Fukao and Paul (2017). 
Note: Sectoral shares in GDP are calculated using real GDP in constant 1934-36 prices for 1874-1940 and constant 2000 prices for 1955-2008. In the current study, the pre-WWII era is from 1885 to1935, and the post-WWII era is from 1955 to 1985. 

To conduct the decomposition analysis, we use seven I-O tables (every 10 years) in the prewar era from 1885 to 1935 and six I-O tables (every 5 years) in the postwar era from 1955 to 1985. These seven sectors include: agriculture, forestry, and fishery; commerce and services; construction;  food;  mining and manufacturing (excluding food and textiles); textiles, and  transport, communication, and utilities.

The results show that the annual growth rate of GDP more than doubled in the post-WWII era compared to the pre-WWII era. The real output growth was the highest in the commerce and services sector throughout the period under study, but there was also rapid growth of output in mining and manufacturing, especially in the second half of the 20th century. Sectoral output growth in mining and manufacturing (textile, food, and the other manufacturing), commerce and services, and transport, communications, and utilities outgrew the pace of growth in GDP in most of the periods. Detailed decomposition results show that in most of the sectors (agriculture, commerce and services, food, textiles, and transport, communication, and utilities), changes in private consumption were the dominant force behind the demand-side explanations. The demand-side effect was strongest in the commerce and services sector.

Overall, demand-side factors — a combination of the Baumol and Engel effects, were the main explanatory factors in the pre-WWII period, whereas  supply-side factors were the key driver of structural transformation in the post-WWII period.

To contact the authors:

Kyoji Fukao,

Saumik Paul,, @saumik78267353


Baumol, William J., “Macroeconomics of unbalanced growth: the anatomy of urban crisis”. American Economic Review 57, (1967) 415–426.

Chenery, Hollis B., Shuntaro Shishido and Tsunehiko Watanabe. “The pattern of Japanese growth, 1914−1954”, Econometrica30 (1962), 1, 98−139.

Fukao, Kyoji and Saumik Paul “The Role of Structural Transformation in Regional Convergence in Japan: 1874-2008.” Institute of Economic Research Discussion Paper No. 665. Tokyo: Institute of Economic Research (2017).

Trade in the Shadow of Power: Japanese Industrial Exports in the Interwar years

By Alejandro Ayuso Díaz and Antonio Tena Junguito (Carlos III University of Madrid)

The history of international trade provides numerous examples of trade in the ‘shadow of power’ (Findlay and O´Rourke 2007). Here we argue that Japanese empire power was as important as factor endowments, preferences, and technology, to the expansion of trade during the interwar years. Following Gardfield et al 2010, the shadow of power that we discuss is based on the use or threat of violence or conquest which depend on the military capabilities of states.

Figure 1:Japan and World Manufacturing Export Performance. Source: Japan and World comparative manufacture exports in volume (1953=100) from UN Historical Trade Statistics.

Japan was a latecomer to 20th-century industrialization, but during the interwar years, and especially in the 1930s, it was able to activate a complex and aggressive industrialization policy to accelerate the modernization of its industry. This policy consisted of import substitution and exports of manufactures to its region of influence. This newly created empire was very efficient in developing a peculiar imperial trade in the shadow of power throughout East and Southeast Asia in conjunction with a more aggressive imperial regional policy through conquest.

The trade generation capacity of the Japanese empire during the interwar years was much higher than that suggested by Mitchener and Weidenmier (2008) for the preceding period (1870-1913). However, some caution needs to be exercised in making this comparison because it might indicate issues associated with the interpretation of the relevant statistics. Japanese empire trade membership increased by more than ten times that associated with the British, German and French Empires, during this period and was twice as great as that for the US and Spanish empires. Consequently, it might be argued that our coefficients are more prominent because they are capturing stronger intra-bloc bias that emerged after the Great Depression.

Employing a granular database consisting of Japanese exports towards 117 countries over 1,135 products at six different benchmarks (1912,1915,1925,1929,1932 and 1938) we are able to demonstrate that the expansion of Japanese exports during the interwar period was facilitated by the exploitation of formal and informal imperial links which exerted a bigger influence on export determination than productivity increases.

Figure 2: Japanese total manufacturing exports by skills and region. Source: Annual Returns of the Foreign Trade of the Empire of Japan.

a) Manufacturing exports by skills

b) high skilled exports by region


The main characteristics of this trade expansion between 1932 and 1938 were high-skill exports directed towards Japanese colonies. Additional evidence indicates that Japan did not enjoy comparative advantage in products with limited export- market potential. Colonial infrastructure, building and urbanization were used as exclusive markets for high-skill exports and became one of the main drivers of Japanese export expansion and its modern industrialization process.

Trade blocs in the interwar years were used as instruments of imperial power to foster exports and as a substitute for productivity in encouraging industrial production. In that sense, Japan’s total exports in 1938 were between 28% and 47% higher than 1912 thanks to imperial mechanisms. The figure is much higher when we capture the imperial effect on high-skill exports (between 66% and 76% higher thanks to imperial connections). The quoted figures are based on a counterfactual comparing exports without the empire to those obtained via Imperial mechanisms.

We believe that our results demonstrate the colonial trade bias mechanism used by imperialist countries was inversely related to productivity. The implicit counterfactual hypothesis would be that without imperial intervention in the region Japan would not have expanded its high-skill exports and would not have exported such a variety of new products. In other words, Japan’s industrialisation process would have been much less pronounced.



Ayuso-Diaz, A. and Tena-Junguito, A. (2019): “Trade in the Shadow of Power: Japanese Industrial Exports in the Interwar years”. Economic History Review (forthcoming).

Findlay, R. and O’Rourke, K. (2007). Power and Plenty. Princeton, NJ: Princeton University Press.

Garfinkel, M, Skaperdas, S., and Syropoulos, C. (2012). ‘Trade in the Shadow of Power’. In Skaperdas, S., and Syropoulos, C. (eds.), Oxford Handbook on the Economics of Peace and Conflict. Oxford University Press.

Mitchener, K. J., & Weidenmier, M. (2008). Trade and empire. The Economic Journal, 118(533), 1805-1834.

Ritschl, A. & Wolf, N. (2003). “Endogeneity of Currency Areas and Trade Blocs: Evidence from the Inter-war Period,” CEPR Discussion Papers 4112.


To contact the authors:

Alejandro Ayuso Díaz (

Antonio Tena Junguito (

Price Shocks in Regional Markets: Japan’s Great Kantō Earthquake of 1923

by Janet Hunter (London School of Economics)

The paper was published on The Economic History Review and is available here on early view 

What do we know about how a market economy operates in the immediate aftermath of a major natural disaster such as an earthquake? Well, actually less than you might think. Specialists in disaster studies have understandably focussed on resilience, relief and reconstruction. The economics of disasters has offered limited frameworks for such addressing this kind of question, although major natural disasters have usefully served in analysis as exogenous shocks or natural experiments. Yet rebuilding economic activity following a major natural disaster must be helped by improving our understanding of the mechanisms whereby the disaster impacts on market activity, and the response of economic actors, both individual and collective. Our work builds on the idea that the analysis of markets in the short-term recuperation phase is best undertaken using the laws of supply and demand, an argument put forward in one of the classic works of the economics of disasters back in 1969.[1]

The Great Kantō Earthquake

The so-called Great Kantō Earthquake of September 1923 in Japan devastated the cities of Tokyo and Yokohama and much of the surrounding area. The location of the disaster is shown on the map.

Map of Japan showing the epicentre of the 1923 earthquake and the cities for which we obtained price data.

While significant damage was caused by the seismic shocks and subsequent tsunami, much of the destruction and casualties (well over 100,000 died) was inflicted by the fires that broke out following the earthquake. The physical destruction was immense, and affected not only the country’s political capital, but its largest urban conglomeration and its major export port. The Kantō plain was also in many respects the hub of an increasingly integrated national economy.

The city of Yokohama, Japan’s most important export port, following the earthquake and fire of September 1923

Our concern in this article has been to analyse the shifts in the availability of, and demand for, different commodities following the 1923 disaster, with a view to identifying the magnitude and duration of such shifts. We have done this through an analysis of price changes for different products over the period before and after the disaster, something that allows us to explore what economists have termed the ‘ripple effects’ of natural disasters.

Our data confirmed that the economic impact of the disaster was far from being confined to the area of destruction. Price changes were experienced across the Japanese archipelago. We did find, though, that the extent of any change tended to diminish the greater the distance from the capital area.

Proportion of the total increase in rice price due to the earthquake by distance

It was also clear that the impact on prices was somewhat stronger in the north and northeastern half of Japan, a region that had traditionally been more closely integrated with Tokyo, than was the case in the southwest, which had long been more closely integrated with the urban area around the city of Osaka. The variation and pattern of the price changes we identified conformed with what we know about patterns of market integration in Japan in the early 20th century. At the same time, the ripples were in many cases less than might have been expected, and certainly less than cotemporary reports suggested. Nor were they in most cases of a significant duration. While there were some initially significant price rises associated with a sudden demand for reconstruction goods, for example, price levels tended to fall again within a few months, suggesting a move back towards some kind of market equilibrium. The pattern of change varied according to the product; a diversity of factors affected supply and demand for different products, and understanding these factors, as well as the pattern of government and institutional intervention in some markets, requires further analysis. It is also the case that analysis of retail prices, for which data are much more difficult to obtain, might well show a somewhat different story from the wholesale price data set that we have been able to compile. Retail prices are, of course, much more difficult to control and a better reflection of what the consumer actually had to cope with.

Overall, though, our analysis confirms that at this time Japan was a relatively well integrated economy. Our findings that prices reverted relatively rapidly toward equilibrium are in line with most other economic indicators showing that there was a relatively rapid reversion to former trends. The disaster, in short, was a short-term exogenous shock from which Japan soon recovered. That does not mean that it did not matter. We still have insufficient knowledge of the factors accelerating or limiting the spread and duration of any price changes following a natural disaster of this kind, and, crucially, of any implications for the longer term consequences of such a shock for the economy as a whole. For contemporary disaster studies, understanding these factors is one of the keys to recovery and the building of resilience.

[1] D.Dacy & H.Kunreuther, Economics of Natural Disasters (New York, 1969).