Copyright 2019 Tony Lima. All Rights Reserved. Copyright © application pending case number 1-768889680. Permission is granted to quote up to 250 words in other articles provided that there is a link to this post near the beginning of your article. For a pdf version of this article, scroll to the bottom.
[Update May 13: It looks like Judy Shelton will be the next nominee. And, sure enough, she supports the gold standard as evidenced in her op-ed in the Wall Street Journal. Dr. Shelton’s Ph.D. is in business administration from the University of Utah. She has taught international finance. Believe me, a Ph.D. in business with an emphasis in economics is very far from a Ph.D. in economics.]
President Trump was considering nominating Stephen Moore and Herman Cain to serve on the Board of Governors of the Federal Reserve system. All three men support a return to the gold standard for U.S. monetary policy. Here’s why they’re dead wrong.
The short version is straightforward. If the quantity of money in circulation is directly tied to the quantity of gold held by the central bank, then it is the growth rate of the gold supply that determines the growth of the money supply. What determines the growth rate of the supply of gold? The only way more gold can be produced is to dig it out of the ground. Would you rather have the growth rate of the money supply determined by a central bank or Barrick Gold (the firm that acquired Homestake Mining in 2001)? While this argument is intuitively appealing, there’s massive empirical support for the proposition that the U.S. economy performed much worse under the gold standard than it has using paper money.
What They Said
Writing in the Journal in 2012, Mr. Cain said,[1] “However imperfect a gold standard may be, it remains the best among all alternatives. The empirical data for both the classical gold standard, which I favor—and even the flawed “gold-exchange” standard, as we had under the Bretton Woods system—are impressive. Economic growth was stronger, unemployment rates lower, the price level more stable, and recessions less frequent and less severe than under the present system.”
He’s actually got it exactly backwards.
The New York Times reported Mr. Moore’s comments:[2] “In the past, Mr. Moore expressed qualified support for a return to the gold standard, which pegs the value of the dollar to gold prices, a practice the United States abandoned in 1971. In an interview in 2015, he agreed with his friend Steve Forbes that a return to such a standard would be ‘a lot better than what we have now.’ ”
And Mr. Trump himself has voiced similar sentiments as quoted in The Hill in 2016.[3] “… Trump openly expresses admiration for the gold standard. ‘We used to have a very, very solid country because it was based on a gold standard,’ he stated during a televised interview last year, three months before announcing his run for president. ‘We don’t have that anymore.’ ”
Caveat
In this article I look at some interesting data. My analysis is in no way complete in that I assumed switching from the gold standard to fiat currency (paper money) was the only important factor influencing the economy. I’ve excluded the period from 1861 to 1950 to remove the Civil War, two World Wars, and the Great Depression. Nevertheless, there are other events that undoubtedly affected economic performance. One that comes to mind immediately is the industrial revolution. Even in 1950 the U.S. economy looked very different from 1850. There are undoubtedly many other factors. Please keep this in mind while you read what follows.
Basic Number Crunching
Before proceeding further, here’s some basic data. I excluded the entire period from 1861 to 1949 because that period included the Civil War, two World Wars, and the Great Depression. B.R. Mitchell compiled annual data on U.S. real GNP since 1800.[4] Between 1800 and 1860 there were seven intervals during which real GNP decreased. That’s well beyond the standard definition of a recession, two consecutive quarters of declining real GDP. The average decline was 14.71%. The deepest recession was 1816-1821. Output declined by a whopping 28.92%. By comparison, between 1929 and 1933 real GNP declined by 46.10%. The Great Depression was really, really bad.
But looking at the period from 1950-2005, there were ten official recessions.[5] The annual percentage change in real GNP was never negative during this period.
Think about that. Output continued to grow through each of those ten recessions. That’s pretty amazing and a testimonial to the vast improvements in understanding how the U.S. economy functions.
Why the Gold Standard Was Bad
Money lubricates the wheels of commerce.[6] In a growing economy the volume of transactions grows. It would be nice if the quantity of money in circulation grew by enough to accommodate this growth.[7] Even Milton Friedman agrees that the money supply should grow over time (although his precise estimate of the growth rate varied quite a bit). If real income grows at 2 percent per year the real money supply should grow at a rate a bit slower than that.[8]
Friedman and others have proposed strict monetary rules: let the money supply grow at two percent per year and ignore everything else. This has the virtue of guaranteeing that inflation will be kept under control. The major defect is that it ignores medium-term variation in the growth rates of income. A good recent example is the “secular stagnation” of slow growth that lasted from 2008 through 2016. If real economic growth is low we want the growth rate of the real money supply to be low. If real economic growth is higher, we want more money in circulation to finance that growth.
But, to a certain extent, none of that matters. If the quantity of money in circulation is directly tied to the quantity of gold held by the central bank, then it is the growth rate of the gold supply that determines the growth of the money supply. What determines the growth rate of the supply of gold?
California, Alaska, Oh My[9]
The only way more gold can be produced is to dig it out of the ground. There were numerous gold strikes during the 19th century, starting with the Carolinas and Georgia. But two created major economic booms. The first was the discovery of gold in California in 1849.[10] The second was the Alaska gold strike of 1880.[11] In 1849 U.S. per capital gross national product grew by -0.03%. The following year, 1850, the growth rate was 13.74%. Similarly in 1880 the growth rate was 16.19%. These bursts of growth were always followed a few years later by a depression. Here’s the data.
(click for larger image)
Academic Research I: Ritschl, et.al.
Economists have studied the gold standard to the point of wondering if there’s anything left to learn. But new techniques are constantly being developed. Here I’ll discuss two recent papers. The first, “The U.S. Business Cycle, 1867-2006: A Dynamic Factor Approach” was written by Albrecht Ritschl, Samad Sarferaz, and Martin Uebele (hereafter “Ritschl et.al.”).[12] The authors use a variant of principal components analysis to aggregate many different variables including a number that are not directly included in the national income accounts. They find that real economic activity, excluding inflation, had the same overall volatility under the “classical” gold standard (1867-1913) and under fiat currency regimes (1946-2006). The period 1914-1945 includes two World Wars and the Great Depression. The authors wisely excluded those 32 years from their study. My guess is that 1914 was chosen because that is the year in which the Federal Reserve was started and paper money was issued.[13] Technically, however, U.S. currency was backed by gold until President Nixon said U.S. dollars would no longer be backed by gold in August, 1971.[14]
In the following description of their results, a range of volatility ratios are reported. The authors have sliced and diced their data many different ways. By my count, they estimate ten different measures of volatility. Two include only non-agricultural activity. Four cover only real data (inflation removed). Four more cover nominal data both overall and non-agricultural. We can learn quite a bit from looking at their results for three different periods: 1979-2006, 1960-1969, and the postwar period 1946-2006. There is one general result that stands out.
Volatility for the nominal series is much lower than the real series. That implies, interestingly, that most stabilization after 1946 took the form of keeping inflation under control. This, of course, flies in the face of those who believe that hard currencies always keep the price level stable. With that in mind, let’s look at specific time periods.
The years following 1979 are described by economic historians as the “Great Moderation.” The exception, of course, is the Great Recession of 2007-2008. And, indeed, the authors report that volatility was considerably lower during this period. They estimate that the volatility of their aggregate economic activity index was between 50% and 75% of volatility in the 1867-1913 era.
Similarly, the 1960s are called the “Golden Age” with high economic growth, low unemployment, and moderate inflation. During this period, volatility ranged between 40% and 65% compared to 1867-1913. The troublesome periods are 1946-1959 and the turbulent 1970s.
Comparing 1946-2006 with the earlier period, trouble arises. Their estimates of volatility for the real economy range from 88% to 125% compared to the pre-1914 period. But their estimates of the nominal series range from 65% to 133%. Three of the four nominal series have volatility ratios below 100%, indicating greater stability in the postwar period. For reference, the authors’ Table 1 is shown below.
(click for larger image)
Academic Research II: Young and Du
Andrew T. Young and Shaoyin Du published “Did Leaving the Gold Standard Tame the Business Cycle? Evidence from NBER Reference Dates and Real GNP” in 2009.[15] Their main concern is identifying the correct year in which the era of longer business cycle expansions began. Other authors use the 1933 departure from the gold standard and the postwar period beginning in 1946.[16]
Young and Du use a Wilcoxon test based on ranking the duration of expansions and recessions by length. They perform a number of tests using various date ranges and methodologies. Their conclusion is pretty definitive:
Considering business cycle durations using NBER reference dates supports the conclusion that the first quarter of 1933 ushered in longer expansions absolutely and relative to adjacent recessions.[17]
But they are also concerned about the volatility of real GNP. They offer this simple statistic:
The standard deviation of the GNP growth rates pre-1950 (III) is 0.0286; almost three times that associated with 1950 (III) onward (0,0108). For the percent deviations from trend series the difference in standard deviations is also dramatic: 0,0512 for pre-1951 (III) and then 0,0187 from that point onward.[18]
Translation: the post-1950 economy is about one-third as volatile as the pre-1950 economy. The question is what event(s) might have caused this increased stability.
One possibility has already been mentioned: the 1946 Full Employment Act. But there was another event more closely related to monetary policy. The Treasury Federal Reserve Accord of March, 1951. Before that the Fed had promised to use monetary policy to stabilize the market for U.S. Treasury securities. This deal released the Fed from that obligation, allowing the Federal Open Market Committee to actually use monetary policy. (Of course, monetary policy during World War II had very different objectives from postwar needs.)
Academic Research: Conclusions
It seems clear that the economy has been more stable since 1950 compared to the period before 1950. Ritschl et.al. show that the reduced volatility was most likely a reduction in the variability of inflation rates. His team used a methodology that includes many more variables than real GNP. But the Young and Du analysis is pretty persuasive. We have good national income data starting in 1929. It appears that the volatility of real GNP has decreased significantly in the post-1950 U.S. My guess is that this is largely due to the Treasury Federal Reserve Accord.
Some Number Crunching
Because I believe in data, I rounded up figures for GNP and population from 1800 to 2005.[19] There is a major problem involving history. The period from 1900 to 1950 included two world wars and the Great Depression. And, of course, the late 19th century featured the Civil War and Reconstruction (1861-1877). Since I am mainly concerned with long-run economic activity during relatively normal economic times I excluded the entire period from 1861 through 1949. For better or worse, this procedure excludes the Alaska gold strikes beginning in 1872. Nevertheless, let’s see if we can make some sense of all this.
First, we need to convert real GNP into real GNP per capita (a better measure of how well off individuals are). Population data was available from Google Fusion.[20] One simple measure is the average rate of growth of real GNP per capita and the standard deviation of the growth rates. Table 1 Shows the results.
(click for larger image)
For a first cut, that’s pretty persuasive. It looks like growth increased and the size of fluctuations decreased in the post-1950 era. But we need to remember that the standard deviation is calculated relative to the mean. There must be some way around this problem.
For the solution I’ll turn to the world of finance. The Sharpe Ratio[21] is used to measure expected return relative to risk. While there are several different forms, at its simplest it is the ratio of the differential rate of return to the standard deviation of the rate of return.[22] As you can see I’ve added the Sharpe Ratio to Table 1. Here’s what the numbers mean.
Think of one standard deviation as one unit of risk. During the first 60 years of the nineteenth century, every additional unit of risk was rewarded with 0.046 units of growth. That’s not very good. In fact, it’s downright bad.
Between 1950 and 2005 every additional unit of risk was rewarded with 1.835 units of growth. That’s much better (to put it mildly).
For those who like their data visual, I inverted the Sharpe ratio and calculated the standard deviation divided by the growth rate in each year. The results are eye-popping.
(click for larger image)
(click for larger image)
Conclusion
After looking at the data and some recent research it’s difficult to argue that returning to the gold standard is a good idea. In fact, I would say it’s a very, very bad idea.
[1] Cain, Herman (2012) “Herman Cain: We Need a Dollar as Good as Gold.” Wall Street Journal, May 13, 2012. Available at https://www.wsj.com/articles/SB10001424052702304070304577395891113592150 accessed May 4, 2019.
[2] Tankersley, Jim (2019) “Trump Taps Fed Critic Stephen Moore for Board Seat.” Available at https://www.nytimes.com/2019/03/22/us/politics/stephen-moore-federal-reserve.html accessed April 10, 2019.
[3] Shelton, Judy (2016) “Trump’s problem with the Fed.” Available at https://thehill.com/blogs/pundits-blog/finance/281683-trumps-problem-with-the-fed accessed April 10, 2019.
[4] Mitchell, B.R. (2007). Data is from the compilation at https://gpih.ucdavis.edu/files/Nominal_GDP_Americas%20_17oct2015.xlsx. Original source is B.R. Mitchell, International Historical Statistics: The Americas, 6th ed., New York: Palgrave Macmillan, 2007; orig. 1983
[5] NBER (2010). “US Business Cycle Expansions and Contractions.” The National Bureau of Economic Research has a Business Cycles Committee that is the semi-official arbiter of when recessions and expansions begin and end. Available at https://www.nber.org/cycles/cyclesmain.html accessed May 3, 2019.
[6] I have no idea who first came up with this idea. I know it is not mine, nor was it invented by the many people who have taught me about money, banking, and monetary theory. If you know the original source, please tell me.
[7] Economists call this the demand for money. People, firms, and other entities want to hold money to facilitate transactions. The quantity of money they want to hold is the total demand for money. The money supply is determined by either the quantity of gold the central bank holds or the quantity of money the central bank chooses to print.
[8] Don’t get me started on the elasticity of the demand for money. Let’s just say the money supply should grow at a slower rate than the real economy because there are economies of scale in the use of money.
[9] Mitchell, B.R. (2007) op.cit.
[10] California state highway 49 transverses gold country. Today it is a delightful tourist destination of small towns, wineries, unusual restaurants, and various attractions. Contact the author for reecommendations.
[11] The Alaska gold strike of 1880 was in Juneau. There were four other major finds: Sitka (1872), Windham Bay (1876), Klondike and Nome (1890), and Anvil Creek (1898). See http://www.alaskascenes.com/alaskagold.html For additional details. Accessed April 18, 2019.
[12] Ritschl , Albrecht, Samad Sarferaz, and Martin Uebele (2016). “The U.S. Business Cycle, 1867-2006: A Dynamic Factor Approach.” The Review of Economics and Statistics, March 2016, 98(1): 159-172.
[13] The Federal Reserve Act of 1913 was the enabling legislation. For more information see https://www.federalreserve.gov/aboutthefed/fract.htm accessed May 2, 2019.
[14] Ghizoni, Sandra Kollen (Federal Reserve Bank of Atlanta, 2013), “Nixon Ends Convertibility of US Dollars to Gold and Announces Wage/Price Controls. August 1971.” Federal Reserve History, November 22, 2013. Available at https://www.federalreservehistory.org/essays/gold_convertibility_ends accessed May 2, 2019.
[15] Young, Andrew T. and Shaoyin Du (2009). “Did Leaving the Gold Standard Tame the Business Cycle? Evidence from NBER Reference Dates and Real GNP.” Southern Economic Journal 2009, 76(2), 310-327
[16] 1946 also saw the passage of the U.S. Employment Act.
[17] Young and Du, op.cit., p.318
[18] Ibid., p.323.
[19] Mitchell, B.R. (2007) op.cit
[20] A note on the Google Fusion site says Google will stop supporting Fusion on December 3, 2019. As of May 4, 2019 the following URL works: https://www.google.com/fusiontables/DataSource?docid=1uq7HDeyUqucF-AMRGMVWsPXwn4cHiEaMCS9E2zYF
[21] Sharpe, William F. (1994). “The Sharpe Ratio.” The Journal of Portfolio Management, Fall 1994. 21(1), pp. 49-58. Available on Prof. Sharpe’s website at https://web.stanford.edu/~wfsharpe/art/sr/sr.htm. Accessed May 4, 2019.
[22] The differential rate of return is the difference between the expected rate of return on a portfolio and the return on a benchmark portfolio, often the rate of return generated by the overall market.
