With protectionist pressures rising and the multilateral trading system seemingly at risk, it is natural to look to the 1930s for evidence of how protection affects the volume and pattern of trade. This column examines the impact of Britain’s decisive break with a longstanding tradition of free trade in 1931, when the country switched dramatically to a policy of protection, discriminating in favor of the British Empire. The shift explains roughly a quarter of Britain’s trade collapse, and around three quarters of the big increase in the British Empire’s share of British imports during the 1930s.
With a protectionist president in the White House, the future of the multilateral, rules-based international trading system seems much less certain than it appeared to be just a few years ago. And the imminent departure of Britain from the European Union – while still not completely certain – risks disintegrating a subset of international markets for goods and services in a way that has not been seen for many years. So it is not surprising that politicians and commentators are turning to the 1930s for examples of what protectionism can imply for international trade flows.
Trade costs and trade flows in the 1930s
We have all heard of the Smoot-Hawley tariff, but the global impact of interwar trade policies remains understudied by economists. Perhaps surprisingly, existing econometric research tends to underplay the quantitative impact of 1930s protectionism. The basic point is straightforward: output and incomes plummeted between 1929 and 1933, and this fact alone can explain the majority of the period’s trade collapse.
The implication is that the increase in trade costs implied by the tariffs and quotas of the period played only a relatively small role in lowering overall trade flows. For example, using aggregate time series evidence, Irwin (1998) estimates that between the second quarter of 1930 and the third quarter of 1932, US imports would have declined by 31.9% even in the absence of any changes in ad valorem tariff rates. The actual decline was 41.2%, indicating that while tariffs mattered, they only accounted for around a quarter of the fall in imports.
World trade not only collapsed during the early 1930s: it also became much less multilateral. Countries like Britain and France, which already had empires, traded more with those empires. For example, the British Empire’s share of British imports rose from 30% to 42% between 1929 and 1938; and the equivalent figures for France were 12% and 26%.
At the same time, countries like Germany and Japan, which were either in the process of or about to begin acquiring empires of their own, similarly traded more intensively with their respective spheres of influence. For example, Japan’s share of imports from its sphere of influence went from 12% to 30% between 1929 and 1938.
Previous researchers have used aggregate data on bilateral trade flows to explore the determinants of these decreasingly multilateral trade patterns. They typically find that the formation of imperial trade blocs (modeled as dummy variables within standard gravity frameworks) were of little or no importance.
But studies relying on aggregate trade flows are problematic. Tariff rates varied greatly across commodities, while the 1930s saw the wholesale introduction of quotas, the impact of which was, by definition, commodity-specific. In addition, trade policy was explicitly discriminatory during the period. This suggests that there may be a substantial benefit to using commodity-level data on trade and trade policy when estimating the impact of the protection of the period.
Britain’s break with free trade
In 1931, Britain broke decisively with a longstanding tradition of free trade. In November, the Abnormal Importations Act allowed the Board of Trade to impose tariffs of up to 100% ad valorem on manufactured goods from outside the Empire, and tariffs of 50% were immediately imposed on many of these.
Another act soon followed allowing for similar duties on non-Empire fruit, flowers, and vegetables. In February 1932, the Import Duties Act imposed a general 10% tariff on goods not already subject to duties, though some important primary imports were exempted.
Finally, the Ottawa conference, which opened in July 1932, saw a series of bilateral trade agreements being signed between Britain and the Dominions. Britain agreed to maintain or raise tariffs imposed on ‘foreign’ (that is, non-Empire) imports under the terms of the 1932 Import Duties Act, and not to reduce the 10% tariff without the consent of the Dominions; to continue to exempt Empire products from these tariffs; and to introduce or enhance Imperial Preference on a wide range of agricultural commodities and raw materials of special interest to the Dominions, by raising duties or by protecting goods that had previously been duty free, such as wheat.
Quotas were also introduced for several agricultural commodities on the basis that policy needed to serve the interests of ‘the home producer first, Empire producers second, and foreign producers last’. The Ottawa accords were thus explicitly discriminatory.
What was the impact of Britain’s dramatic policy shift?
To answer this question, we collected detailed, commodity-level data on bilateral trade flows, tariffs, and non-tariff barriers to trade, all of which varied at the commodity-country-year level. This involved typing information on British imports of 847 individual product categories coming from 49 countries and sub-regions; the end result was a dataset giving imports into Britain of 258 consistently defined commodities from 42 countries over the 15-year period 1924-38.
We also collected information on tariffs, quotas, voluntary export restraints, and other variables potentially influencing trade flows during the period, such as the presence or absence of international cartels. To quantify the impact of the 1931-32 switch to protection, we compared the trade flows that were actually observed from 1931 onwards with the counterfactual flows that would have been observed had tariffs and quotas remained unchanged between 1930 and 1938.
Calculating counterfactual trade flows requires a model. We assumed what are known as nested constant elasticity of substitution utility functions, as in Broda and Weinstein (2006) – see Figure 1.
Following their analysis, we refer to imports of a particular good from a particular country as a variety. At the top level, the representative consumer chooses between domestically produced output and imports. At the second level, the consumer chooses between imports of our 258 goods. And at the third level, the consumer chooses between the 42 potential varieties of each good.
On the supply side, a single production sector transforms the sole factor of production into either domestically produced (and consumed) output or exports via a constant elasticity of transformation production function, as in Anderson and Neary (1996).
The model can be calibrated using information on just GDP and commodity and country-specific imports, as well as four types of elasticities: the top-level elasticity of substitution between domestically produced output and imports κ; the intermediate-level elasticity of substitution between our 258 goods γ; the lower-level elasticities of substitution between different varieties of each good σg; and the supply-side elasticity of transformation η. We estimated the demand-side elasticities, and took estimates of the supply-side elasticity from previous research.
Because the elasticities were (mostly) estimated econometrically, we wanted to take account of the fact that they were imprecisely measured. When performing counterfactual analysis, we therefore drew 1,000 elasticity values from normal distributions, the means of which were the point estimates of the elasticity in question, and the standard deviations of which were the standard errors. We then solved our model to generate the counterfactual ‘constant 1930 trade policy’ equilibrium 1,000 times.
The changing value and geographical composition of British imports
Figure 2 shows the percentage impact of the shift to protection after 1930 on the total value of British imports. It plots not only the average impact across our 1,000 repetitions for each year, but also the 5th, 25th, 75th, and 95th percentiles.
The shift towards protection reduced the value of British imports by 9-10% on average, with the biggest impact being felt in 1933. In that year the mean estimated impact was 10.8%, with 25th and 75th percentile impacts of 8.1% and 13.5% respectively. Protection accounted for about a quarter of the total decline in British imports, which is consistent with the results for the United States cited earlier.
Figure 3 plots the actual share of British imports coming from the British Empire, as well as the counterfactual share that would have obtained had Britain not shifted towards protection after 1930. Again, the figure plots not only the mean counterfactual share, but also the 5th, 25th, 75th, and 95th percentiles.
As can be seen, the shift towards protection had a big effect on the geographical composition of British imports. For example, the Empire’s share of British imports rose from 27% to 39.2% between 1930 and 1935. In the absence of protection it would only have increased to 31.4% according to our mean estimate. Our mean estimate implies that protection accounted for 77% of the shift towards Empire between 1930 and 1933 (the 25th and 75th percentile equivalents being 67.7 and 86.1% respectively). These are big effects.
Our research shows that using disaggregated data does not significantly change the estimated impact of protection on the total value of trade; it does, however, matter a lot for the estimated impact of protection on the geographical composition of trade. Studies using aggregate data find that imperial trade blocs did not have a big influence on trade patterns during the 1930s; in contrast, we find that British trade policy was crucial in increasing the share of the British Empire in British imports.
This ultimately reflects the large estimates obtained for the elasticities of substitution between different varieties of the same good, as well as the extent to which British trade policy after 1930 was discriminatory. Methodologically, our results suggest that there are substantial advantages to using disaggregated trade and trade policy data, and to studying in detail the tariff and non-tariff policies actually pursued by trade bloc members (as opposed to simply looking at whether or not such blocs exist).
The wider impact of interwar protection
Some economists might be tempted to infer from our results that protection did not matter very much for welfare during the period. The very fact that different varieties of the same goods were close substitutes for each other meant that British importers were able to switch from ‘foreign’ goods to ‘British’ goods coming from the Empire, thus mitigating the welfare loss.
But that would be to miss the point. In the context of the 1930s, the ‘Balkanization’ of international trade not only reflected, but probably also exacerbated, the international tensions of the period. That is certainly what contemporary observers thought.
For example, according to Hilgerdt (1935), ‘As bilateralism particularly renders the supply of raw materials to certain countries difficult, it threatens to lead to an intensified fight for influence upon (or the domination of) the undeveloped countries, and thereby to political controversies, which may adversely affect all forms of peaceful collaboration between nations’.
During the war, Condliffe (1941) wrote that ‘it is now so obvious as to hardly need statement that bilateral trade took on aggressive and destructive aspects as international rivalries were sharpened in the era of what is now known as pre-belligerent’).
And so it is no surprise that the 1941 Anglo-American Atlantic Charter committed Britain and the United States to endeavoring after the war, ‘with due respect for their existing obligations, to further the enjoyment by all States, great or small, victor or vanquished, of access, on equal terms, to the trade and to the raw materials of the world which are needed for their economic prosperity’,
Nor is it a surprise that Article 1 of the GATT banned discriminatory trade policies (subject to a number of well-known exceptions). Non-discriminatory trade policies are economically efficient, but the interwar period teaches us that their importance goes well beyond economics. Those lessons remain relevant today.
This article summarizes ‘When Britain Turned Inward: The Impact of Interwar British Protection’ by Alan de Bromhead,, Alan Fernihough, Markus Lampe and Kevin Hjortshøj O’Rourke, which is forthcoming in the American Economic Review.
Alan de Bromhead and Alan Fernihough are at Queens University Belfast. Markus Lampe is at the Vienna University of Economics and Business. Kevin Hjortshøj O’Rourke is at All Souls College, Oxford.