Out of Control: Recognizing Instability


This graph1 shows the ratio of the Dow Jones Industrial Average (DJIA) index to the price of gold.  While the DJIA is rightly criticized as an imperfect indicator of economic activity, it does represent claims on wealth produced by the largest and most successful firms in the US economy. In a free market, one would expect the ratio to trend upwards, as the market creates new wealth while the amount of gold remains nearly fixed. From 1900 until about 1920 this is exactly what happened, with relatively small excursions around a slight upward trend. Beginning around 1920 the trend is replaced by ever-growing gyrations representing the creation and destruction of wealth, enterprises, and livelihoods. Engineers familiar with control theory will recognize the growing oscillations as the onset of instability. 

Engineers have been studying control theory since the 1860s. Keeping any machine operating within narrow boundaries can be difficult. While the mathematics can be complex, many of the lessons learned are familiar to anyone who drives a car.

When driving a car along a straight and narrow roadway the driver cannot simply hold the steering wheel fixed; small imperfections in the road and other perturbations cause the car to depart from the desired path. When the car drifts to the left, the driver must steer slightly to the right, and when the car goes right the driver steers left. The steering corrections are proportional to the deviation; if the car is drifting slowly, the correction is small, but if the car is rapidly departing the road, the steering changes must be larger.

This proportional correction of deviations is called negative feedback. The feedback is negative because it is opposite to the error; steering left when the car veers right.

Positive feedback occurs when the correction is in the same direction as the error. If the car veers left, steer left. The car will rapidly leave the road and crash. Not all positive feedback is bad; used to reinforce desired behaviors, such as praising a pet or rewarding an employee for a job well done, positive feedback tends to produce more of the same. This holds true when positive feedback is applied to undesired outcomes; the system will produce still more. Positive feedback acts to amplify the effects of initially small perturbations until they become very large, or some other limit comes into play. The Dow/gold graph shows one system that is clearly subject to ever growing swings.

There are other ways to destabilize controlled systems, even with negative feedback. If there is too much delay between deviation and correction, the deviations tend to grow larger before they are corrected. Drunk drivers weave back and forth in part because alcohol slows reaction time, introducing a lag in the driver's control of the car.

Another way to destabilize a system is with excessive negative feedback. While the correction remains proportional to the deviation, over-correction can lead to a loss of control. If the driver makes violent corrections to small errors, the car will lurch back and forth, eventually spinning out of control and crashing.

There are examples of all three destabilizing mechanisms in recent Fed interventions in the marketplace. Before the creation of the Federal Reserve, banks that made poor decisions about loans went bankrupt and were closed. The market reallocated any remaining capital to better and higher uses. This is negative feedback in action. After creation of the Fed to support failing banks and particularly after creation of Federal deposit "insurance" released depositors from the consequences of patronizing poorly run banks, this negative feedback was largely eliminated.

Instead, we now see Bernanke and his co-conspirators furiously creating money to prop up failing financial firms. This is positive feedback, and it will certainly produce more failing firms and bad debts. Delay is also evident, as the crisis was manifest this past July and August, but the Fed didn’t react until months later.  Jim Cramer's infamous rant had one thing right: the Fed had no idea how bad things were. Cramer started his early August rant by discussing problems at Bear Sterns; the fed approved a loan to support Bear Sterns in mid-March, over 7 months after the onset of the crisis and too late to save the firm.

Meanwhile Bush and Congress lard on more debt for an “economic stimulus package” when any prudent enterprise would be retrenching and saving.  Checks are being sent to the same Americans who took on mortgages and other forms of debt that they cannot afford to repay. This is more positive feedback; excessive debt and credit caused this panic, more debt cannot correct it and will make it worse. The “stimulus” is deliberately targeted to avoid giving money to the most productive citizens, with no checks for Americans in the top 10% or so of filers who pay 70% of the income taxes. The checks won’t be mailed for several months, so we have positive feedback with delay.

It's hard to know if the Fed's actions are excessive, but they are certainly growing rapidly in amplitude. The August injections of a few billion dollars have now grown to multiple efforts that will each create hundreds of billions of dollars, and probably trillions in new debt and credit.

Correcting a credit crisis caused by excessive debt and profligate spending requires three kinds of negative feedback: sound money, saving, and budget cutting. This is the opposite of our present course of money creation, more debt, and more spending. The instability is growing because of this positive feedback.

It was not always so. Murray Rothbard's outstanding treatment of The Panic of 1819 documents striking similarities to the Panic of 2008. There was speculation in real estate fueled by over-easy granting of credit, a loss of confidence in banks and financial institutions, a wave of bankruptcies, and a full-blown credit crisis with a sudden loss of liquidity. Public land sales dropped over 90%, from $13.6 million in 1818 to $1.3 million in 1821. Those favoring government intervention proposed various forms of debtor relief, tighter regulation of banks, inflation of the money supply, and protective tariffs.

While today's media, politicians, bankers, and regulators are nearly unanimous in their calls for more aggressive creation of debt by the Fed coupled with more regulation and spending by the government, during the Panic of 1819 nearly the opposite was true. President Monroe largely ignored the crisis, mentioning the resulting short depression only in explanation of declining federal revenues. Newspapers and magazines of the day tended to support a laissez-faire approach. The New York Evening Post wrote that "Time and the laws of trade will restore things to an equilibrium, if legislatures do not rashly interfere to the natural course of events."2

Despite the many similarities in causes and effects between the two Panics, opinions in today's "main stream" media are directly opposite to the wisdom of 1819, although the internet provides a welcome source of opposing views for those interested in finding them. Among politicians, Congressman Ron Paul stands nearly alone in his staunch support of limited government, less regulation, drastically reduced government taxation and spending, and a return to sound money.

The correction for creating excessive amounts of money out of thin air is automatically provided by the gold standard, as gold cannot be created and is mined only with great difficulty and expense. Our nation grew and prospered under various forms of gold-backed money for 125 years. In the 37 years since the federal government abandoned any pretense of gold backing of our currency our standard of living has declined and our once mighty economy has been reduced to a hollow shell.

The tie between the dollar and gold was damaged by creation of the fed in 1913 and progressively weakened thereafter until being abandoned completely in 1971. The graph shows that the removal of the last tie between gold and the dollar caused a rapid fall in the Dow/gold ratio as the price of gold increased. The market had to find the correct price for gold after decades of a government-fixed price despite an ever-growing flood of newly created dollars. Once the initial correction and over-reaction (another common trait in feedback systems) for the market price of gold had passed, the fed was free to further inflate the money supply and start the latest and greatest excursion in the Dow/gold ratio.

Paying down debts and increasing savings require determination and willpower. These measures can and should be adopted by individuals, businesses, and governments. As Jane Austen wrote in Persuasion, "Your debts are extreme. You must retrench." Cutting spending for individuals and private business is straightforward; once the money and credit run out, spending falls automatically if it hasn't been cut voluntarily. Governments with the power to create unlimited money and credit can escape these limits, but neither governments nor their subjects can escape the consequences of such foolish behavior.

It will take all three of these forms of negative feedback, applied without delay, to correct the growing instability. The alternative is nearly unthinkable, but history provides plenty of examples. Inappropriate feedback and excessive delays have doomed many man-made systems. The awesome pyramid of debt and credit enabled by the Fed’s machinations appears to be headed for the same “crash and burn” fate. The free market produces countless feedback mechanisms that tend to keep economies growing, creating more wealth and better living standards for everyone. Apparently Ben Bernanke never took advantage of the excellent but demanding courses in control systems engineering offered by MIT while completing his Ph.D. in economics.  I don’t believe that the market can or should be subject to central control, but if I were to take the job of controlling the money supply, I would want a solid background in control theory.


  1. My graph was inspired by a similar plot produced by Nick Laird that shows a long period of relatively stable growth from 1800–1920. Since the DJIA was first published in 1896, Mr. Laird uses a proxy for the DJIA 1800–1896. My graph was constructed using monthly average price data provided by the Dow Jones Indexes and Wren Research. I used government-fixed gold values of $20.67 per ounce for the period 1900-1934, and $35 from 1934–1968. Kitco provided monthly average prices for December 2007 to March 2008. The last point in the graph averages price data from March 1–14, 2008.
  2. New York Evening Post, June 15, 1819, quoted in The Panic of 1819: Reactions and Policies, Murray N. Rothbard, Ludwig von Mises Institute, 2007, pp 31–32.

March 22, 2008