Friday, December 9, 2011

The Depression of 1920–1921: An Austrian Myth

The Austrians and other libertarians frequently assert that America had a depression from 1920 to 1921. That is simply not true if one accepts the estimates of (1) Romer or (2) Balke and Gordon.

The generally accepted definition of a depression is a real GNP/GDP contraction of 10% or more - or, in other words, where the real (inflation-adjusted) value of national output falls by 10% or more. In past estimates of the fall in national output, such as the official Commerce Department data (based on the Kuznets-Kendrick series), GNP fell 8% between 1919 and 1920 and 7% between 1920 and 1921 (Romer 1988: 108). But that data has been challenged, and acceptance of it also shows numerous and long depressions in the 19th century, depressions that did not end quickly – data which hardly supports the Austrian apologetics for gold standard capitalism. In fact, if the Austrians really wish to accept the data from the Kuznets-Kendrick series, then that entails that gold standard capitalism was hit by numerous and frequent recessions and depressions, and sometimes very long ones. Other Austrians are of course aware of the most recent GNP estimates, and press these into service to show that the 19th century US business cycle was less volatile than in Kuznets-Kendrick or the Commerce Department data. But you cannot have it both ways: either (1) the Kuznets-Kendrick series is valid or (2) better data is available in Romer (1988) or Balke and Gordon (1989).

What do the most recent GNP estimates show?

First, we can cite Romer’s estimates for GNP here (I have added the annual growth rates by my own calculation):
Year | GNP* | Growth Rate
1914 | $414.599
1915 | $443.048 | 6.86%
1916 | $476.498 | 7.54%
1917 | $473.896 | -0.54%
1918 | $498.458 | 5.18%
1919 | $503.873 | 1.08%
1920 | $498.132 | -1.13%
1921 | $486.377 | -2.35%

1922 | $514.949 | 5.87%
1923 | $583.105 | 13.23%
* Billions of 1982 dollars
(Romer 1989: 23).
The figures show a GNP contraction of 3.47% from 1919 to 1921, a recession that was less severe than the contraction of 1908 by Romer’s figures, and much lower than the 10% figure needed to declare a depression. Romer concludes that the “growth path of output was hardly impeded by the recession” (Romer 1988: 108–112), and the positive supply shocks that resulted from the resumption in international trade after WWI actually benefited certain sectors of the US economy (Romer 1988: 111).

Secondly, the estimates for GNP of Balke and Gordon are here (I have added the annual growth rates by my own calculation):
Year | GNP* | Growth Rate
1914 | $402.4 |
1915 | $417.3 | 3.70%
1916 | $485.0 | 16.2%
1917 | $484.9 | -0.02%
1918 | $522.2 | 7.69%
1919 | $507.1 | -2.89%
1920 | $496.3 | -2.12%
1921 | $478.8 | -3.52%

1922 | $513.2 | 7.18%
1923 | $585.0 | 13.99%
1924 | $600.5 | 2.64%
* Billions of 1982 dollars
(Balke and Gordon 1989: 84–85).
These figures show a GNP decline of 5.58% from 1920–1921. That was a severe recession, but hardly a depression.

If we include the contraction at the end of First World War in 1919, the GNP contraction was 8.31% from the wartime GNP peak of 1918 to 1921, but that figure is misleading, as it includes the contraction that appears in Balke and Gordon’s estimates that was caused by the dismantling of America’s wartime economy. Curiously, no such contraction for 1919 appears in Romer’s data. According to the offical data the business cycle ran through these phases in these years:
August 1918–March 1919, severe depression from the end of wartime production;
April 1919-December 1919, expansion;
January 1920-July 1921, recession.

US Business Cycle Expansions and Contractions, National Bureau of Economic Research.
All in all, there was no depression in 1920–1921. There was no financial crisis, and no mass bank failures as in 1929–1933. This was a recession of an entirely different order from that of 1893–1894, 1907–1908 or 1929–1933 (a useful paper on the differences between the 1920-1921 and 1929-1933 contractions in the neoclassical literature can be found in Gertler 2000, a response to Cole and Ohanian, 2000).

BIBLIOGRAPHY

Balke, N. S., and R. J. Gordon, 1989. “The Estimation of Prewar Gross National Product: Methodology and New Evidence,” Journal of Political Economy 97.1: 38–92.

Cole, H. L. and L.E. Ohanian, 2000. “Re-examining the Contributions of Money and Banking Shocks to the U.S. Great Depression,” in B. S. Bernanke and K. Rogoff (eds.), NBER Macroeconomics Annual 2000, MIT Press, Cambridge, MA. 183-226.

Gertler, M. 2000. “Comment,” in B. S. Bernanke and K. Rogoff (eds.), NBER Macroeconomics Annual 2000, MIT Press, Cambridge, MA. 237-257.

Romer, C. D. 1989. “The Prewar Business Cycle Reconsidered: New Estimates of Gross National Product, 1869–1908,” Journal of Political Economy 97.1: 1–37.

9 comments:

  1. Austrians don't define depressions as contractions of 10% or more, so you can't say that they are wrong to call the 1920-1921 period a depression. You can only say they are using a different definition.

    Austrians consider recessions and depressions as essentially identical, because Austrians KNOW that these times are times of correction needed after previous monetary manipulation. Austrians call them "bust" periods. Now, depending on how much the government interferes, the bust's length could differ. The more the government tries to reverse it with money manipulation, the longer it will last. The less the government tries to reverse it with money manipulation, the shorter it will last.

    Austrians argue that the bust of 1920-1921 did not become a protracted period of correction, because the government was still "liquidationist." At that was the last time they were.

    You're just playing a pedant's semantics by saying that there was no depression in 1920-1921 by saying "Wrong, because the definition of depression is..."

    The Dow Jones Industrial Average reached a peak of 119.6 on November 3, 1919. The market bottomed on August 24, 1921, at 63.9, a decline of 47%.

    Unemployment peaked at almost 9% in 1921.

    http://i.imgur.com/vkCgu.gif

    Regardless of what you CALL it, the events were CONSISTENT with Austrian theory.

    http://en.wikipedia.org/wiki/Depression_of_1920%E2%80%9321

    ReplyDelete
  2. "Austrians consider recessions and depressions as essentially identical, because Austrians KNOW that these times are times of correction needed after previous monetary manipulation. Austrians call them "bust" periods. "

    In which cause, the US was hit by a "depressions" in the 1800s frequently, even when it had no central bank. In the 1890s, there was a severe one which caused very high unemployment that persisted for nearly a decade. There was no swift recovery under American gold standard capitalism to low unemployment. So much for the system that was supposed to be superior to (1) a system with a central bank or (2) a central bank and Keynesian demand management.

    Austrians argue that the bust of 1920-1921 did not become a protracted period of correction,

    Protracted!? The recession lasted from January 1920 to July 1921 (a period of 18 months). A recession lasting 18 months is in fact a very long one by the standards of the post-1945 US business cycle.

    By the standard data, the average duration of recessions in peacetime from 1854 to 1919 was 22 months (Knoop 2010: 13), and the average duration of recessions from 1919 to 1945 was 18 months (Knoop 2010: 13). This 1920-1921 bust wasn't even short by 19th century standards.

    ReplyDelete
  3. "Austrians consider recessions and depressions as essentially identical, because Austrians KNOW that these times are times of correction needed after previous monetary manipulation. Austrians call them "bust" periods."

    In which cause, the US was hit by a "depressions" in the 1800s frequently, even when it had no central bank.

    Austrians hold that it's not central banking per se that causes the business cycle. It's the credit expansion specifically. It's just that in our society today, central banks are the main facilitators of it, because they constantly top up bank reserves to enable to them to continue to expand credit.

    Banks expanded credit in the 1800s and moved away from 100% reserve banking on more than one occasion. And the recessions were far shorter in duration than post 1930s recessions, once Keynesianism took hold and delayed the necessary corrections.

    In the 1890s, there was a severe one which caused very high unemployment that persisted for nearly a decade.

    False. Real GNP for the 1890s, according to Romer:

    1890 $182.964 4.53%
    1891 $191.757 4.80%
    1892 $204.279 6.53%
    1893 $202.616 -0.81%
    1894 $200.819 -0.88%
    1895 $215.668 7.39%
    1896 $221.438 2.67%
    1897 $233.655 5.51%
    1898 $241.459 3.33%
    1899 $254.728 5.49%

    I the 1890s, there was only a mild recession in 1893-94. Every other year had spectacular real growth.

    Unemployment is simply not a measure of economic health. An economy whereby everyone is digging ditches and there is no unemployment, is better than an economy whereby there is 10% unemployment but real productivity is increasing. For unemployed people can be supported while production expands in a growing economy, whereas people who are employed in a stagnating economy aren't increasing their standards of living.

    What ultimately matters is people's standard of living, not whether they have jobs or not.

    ReplyDelete
  4. There was no swift recovery under American gold standard capitalism to low unemployment.

    Swift compared to the Great Depression and today's depression, the times when Keynesian responses were at their maximums.

    Austrians argue that the bust of 1920-1921 did not become a protracted period of correction,

    Protracted!? The recession lasted from January 1920 to July 1921 (a period of 18 months).

    Yes. Compare that to the Great Depression of 15 years, and the current depression of going on 3 years now and most likely will be extended into the future precisely because the government won't let the painful corrections occur, and it wouldn't be off the mark to argue that if the government and Fed adopted a 1920-1921 like response, we'd be out of the doldrums, instead of lingering in them like we are now.

    A recession lasting 18 months is in fact a very long one by the standards of the post-1945 US business cycle.

    By the standard data, the average duration of recessions in peacetime from 1854 to 1919 was 22 months (Knoop 2010: 13), and the average duration of recessions from 1919 to 1945 was 18 months (Knoop 2010: 13).

    That contradicts the data from Romer

    1869 75.609
    1870 $76.464 1.13%
    1871 $76.952 0.638%
    1872 $89.605 16.4%
    1873 $94.863 5.86%
    1874 $96.205 1.414%
    1875 $97.684 1.53%
    1876 $104.628 7.10%
    1877 $110.797 5.89%
    1878 $118.906 7.31%
    1879 $127.675 7.37%
    1880 $139.990 9.64%
    1881 $143.580 2.56%
    1882 $149.307 3.98%
    1883 $152.097 1.86%
    1884 $155.684 2.35%
    1885 $157.789 1.35%
    1886 $164.375 4.17%
    1887 $169.453 3.08%
    1888 $168.940 -0.3%
    1889 $175.030 3.60%
    1890 $182.964 4.53%
    1891 $191.757 4.80%
    1892 $204.279 6.53%
    1893 $202.616 -0.81%
    1894 $200.819 -0.88%
    1895 $215.668 7.39%
    1896 $221.438 2.67%
    1897 $233.655 5.51%
    1898 $241.459 3.33%
    1899 $254.728 5.49%
    1900 $264.540 3.85%
    1901 $284.908 7.69%
    1902 $291.572 2.33%
    1903 $306.239 5.03%
    1904 $307.127 0.28%
    1905 $323.162 5.22%
    1906 $351.499 8.76%
    1907 $361.920 2.96%
    1908 $346.800 -4.17%
    1909 $368.872 6.36%
    1910 $383.888 4.07%
    1911 $391.858 2.07%
    1912 $407.112 3.89%
    1913 $424.492 4.26%

    Where are the 22 months long declines?

    This 1920-1921 bust wasn't even short by 19th century standards.

    Well, the larger and longer the prior credit expansion (WW1 was financed by a very large quantity of inflation and credit expansion), the longer it takes to fix.

    ReplyDelete
  5. "An economy whereby everyone is digging ditches and there is no unemployment, is better than an economy whereby there is 10% unemployment but real productivity is increasing."

    Should read

    "An economy whereby everyone is digging ditches and there is no unemployment, is not necessarily better than an economy whereby there is 10% unemployment but real productivity is increasing."

    ReplyDelete
  6. "Compare that to the Great Depression of 15 years, and the current depression of going on 3 years now"

    That is just rubbish. Typical dishonesty.

    (1) You claim that there was a depression of 15 years even though from 1933-1937 and 1938-1940 there was high real GNP/GDP growth and falling unemployment

    (2) you claim that 2008-2011 is a "depression" even though America has had positive real GNP/GDP growth rates since 2009.

    (2) By the same dishonest idiocy I can claim that the whole period 1893-1900 was a depression too.

    ReplyDelete
  7. "A recession lasting 18 months is in fact a very long one by the standards of the post-1945 US business cycle."

    You have not answered that correct statement.

    "Where are the 22 months long declines?"

    Of course it is well known that Romer's data shows less output volatility, but if you accept her data 1920-1921 was just a minor reccession, yet went on for a long time relative to the length of recessions post-1945.

    Anyway, 1893-1894 shows a recession that looks close to 22 months or in that range.

    ReplyDelete
  8. "Real GNP for the 1890s, according to Romer:"

    Real GNP for the 1929-1940:

    Year | GDP* | Growth Rate
    1929 | $977,000
    1930 | 892,800 | -8.61%
    1931 | $834,900 | -6.48%
    1932 | $725,800 | -13.06%
    1933 | $716,400 | -1.29%
    1934 | $794,400 | 10.88%
    1935 | $865,000 | 8.88%
    1936 | $977,900 | 13.05%
    1937 | $1,028,000 | 5.12%
    1938 | $992,600 | -3.44%
    1939 | $1,072,800 | 8.07%
    1940 | $1,166,900 | 8.77%

    Outside of 1929-1933 and 1937-1938, every other year had spectacular real growth: so
    you write rubbish above about a 15 year depression here.

    Yet there is obvious there was a major problem in these years: unemployment.

    Either high unemployment is a measure of the helath of an economy or it is not.

    ReplyDelete
  9. "And the recessions were far shorter in duration than post 1930s recessions"

    (1) Romer's annualised GNP figures, as those of even Balke and Gordon, can conceal recessions lasting 2 or 3 quarters, if growth in the last 2 quarters /1 quarter was strong.

    (2) Romer's annualised GNP figures, as those of even Balke and Gordon, can conceal recessions spanning 2 years if overall growth in both these years was positive. Therefore unless you have reliable quarterly data from the 19th century, you don't have a proper comparison with post 1945 data.

    Anyway, 1945-1973 GNP:

    Year | GNP* | Growth Rate
    1946 | 1798.20 |
    1947 | 1784.80 | -0.74%
    1948 | 1864.80 | 4.48%
    1949 | 1854.20 | -0.56%
    1950 | 2016.50 | 8.75%
    1951 | 2174.30 | 7.82%
    1952 | 2257.40 | 3.82%
    1953 | 2360.10 | 4.54%
    1954 | 2346.20 | -0.58%
    1955 | 2515.80 | 7.22%
    1956 | 2566.90 | 2.03%
    1957 | 2619.20 | 2.03%
    1958 | 2592.90 | -1.00%
    1959 | 2778.10 | 7.14%
    1960 | 2848.20 | 2.52%
    1961 | 2916.10 | 2.38%
    1962 | 3094.10 | 6.10%
    1963 | 3230.10 | 4.39%
    1964 | 3417.50 | 5.80%
    1965 | 3636.40 | 6.40%
    1966 | 3869.80 | 6.41%
    1967 | 3967.70 | 2.52%
    1968 | 4160.60 | 4.86%
    1969 | 4288.00 | 3.06%
    1970 | 4295.80 | 0.18%
    1971 | 4442.20 | 3.40%
    1972 | 4678.90 | 5.32%
    1973 | 4960.30 | 6.01%

    (1) The 1945-1947 contraction is merely postwar conversion.

    (2) Removing 1945-1947, you have only 3 recessions showing up in the data over 26 years.

    ReplyDelete