Under Appreciated Article on the Classical Gold Standard

Bordo and Redish published "Is Deflation Depressing? Evidence from the Classical Gold Standard" in 2003. I have not seen it cited in my research. Their findings are under appreciated. They conclude that only changes in aggregate demand, particularly changes in demand for gold, had a long run impact on the price level during the classical gold standard, although increase in output had short run effects on the price level. They also find that demand shocks did not significantly impact output.

We distinguish between good and bad deflations. In the former case, falling prices may be caused by aggregate supply (possibly driven by technology advances) increasing more rapidly than aggregate demand. In the latter case, declines in aggregate demand outpace any expansion in aggregate supply. This was the experience in the Great Depression (1929-33), the recession of 1919-21, and may be the case in Japan today. In this paper we focus on the price level and growth experience of the United States and Canada, 1870-1913. Both countries adhered to the international gold standard. This meant that the domestic price level was largely determined by international (exogenous) forces. In addition, neither country had a central bank which could intervene in the gold market to shield the domestic economy from external conditions. We proceed by identifying separate supply' shocks, money supply shocks and demand shocks using a Blanchard-Quah methodology. We model the economy as a small open economy on the gold standard and identify the shocks by imposing long run restrictions on the impact of the shocks and on output prices. We then do a historical decomposition to examine the impact of each shock on output. The results for the U.S. are clear: the different rates of change in the price levels before and after 1890 are attributed to different monetary shocks, but these shocks explain very little of output growth or volatility, which is almost entirely a response to supply' shocks. For Canada the results are murkier. As in the U.S., the money supply shocks before 1896 are predominantly negative and after that are largely positive. However, they are non-neutral, and relative to the U.S., money supply shocks play a larger role in determining output behavior in Canada. The key conclusion of our analysis is that the simple demarcation of good vs. bad deflation, where either prices fall because of a positive supply shock, or prices fall because of a negative demand (money) shock does not capture the complexity of the historical experience of the pre-1896 period. Indeed, we find that prices fell as a result of a combination of negative money supply shocks and positive supply shocks.

A Note on Institutions as Aids to Intertemporal Division of Labor and Guarantors of Liberty

Long lived institutions promote an intertemporal division of labor. Our forbearers solved particular problems. Institutions allows us in the present to rely on the wisdom of the past concerning these problems and work on more pressing issues. Hayek notes in The Constitution of Liberty that we do not, “as Thomas Jefferson believed… ‘ascribe to the men of the preceding age a wisdom more than human, and… suppose what they did to be beyond amendment.” There is here a confusion between the wisdom of a particular person and the collective wisdom of persons across many generations:

Far from assuming that those who created the institutions were wiser than we are, the evolutionary view is based on the insight that the result of experimentation of many generations may embody more experience than any one man possesses.

The formation of institutions across time represent the creation and application of information to relevant social problems. These solutions, as embodied in common law, have tended to increase freedom of the individual. The weakening of these institutions by increased reliance on legislation diminishes the freedom procured by this knowledge. It is this danger that Hayek refers to in his critique of rationalism. If the state is empowered to make too swift of changes too often and to deviate from the principles that underlie a liberal society, those principles, which secure the freedom to engage in any action not restricted, may be lost:

Not only is liberty a system under which all government action is guided by principles, but it is an ideal that will not be preserved unless it is itself accepted as an overriding principle governing all particular acts of legislation. Where no such fundamental rule is stubbornly adhered to as an ultimate ideal about which there must be no compromise for the sake of material advantages – as an ideal which, even though it may have to be temporarily infringed during a passing emergency, must for the basis of all permanent arrangements – freedom is almost certain to be destroyed by piecemeal encroachments.

It is on this intellectual foundation that one might argue for an encumbering of legislative action and increased reliance on common law as a means for promoting the general welfare.

Hayek and Foucault: Two Views On Institutions and Coercion

Perhaps the prime problem that society faces is that of violence. Every individual is capable of violence. And even if 99 percent of individuals do not engage themselves initiate violence, the actions of the few that do employ violence, if left unchecked, can be terminally destabilizing for society.

The hallmark of modern society is the existence of robust institutions that constrain violence, promoting its predictability. For example, I know that if I attempt to steal a vehicle or illegally enter a residence that I can expect a violent response from either the owner, local security, or the police. If I don’t pay my taxes, I can expect to end up in court and maybe have an unpleasant interaction with armed agents from the IRS. We are constantly confronted by implicit threats of violence that, at least in part, guide our actions. Violence is the norm, not the exception, for society. Even when we have overcome violence, it is overcome by threat of violence embodied in particular institutions – i.e., property rights, civil liberties. These rights represent protection from force, protection that is organized according to particular rules. With violence as a given, how do we move toward a state of freedom?

This is an issue that both F.A. Hayek and Michel Foucault confront. The former is optimistic about the range of institutional arrangements that might secure liberty. The latter views legal tradition as an extension of warfare, and therefore as a system of oppression. For Foucault, institutional evolution is an extension of oppression that might be one day overthrown. Though in ways opposed, the views of these authors concerning institutions also inform one another.

In The Constitution of Liberty, Hayek’s first task is to define liberty in light of the problem of violence:

It is often objected that our concept of liberty is negative. This is true in the sense that peace is also a negative concept or that security or quiet or the absence of any particular impediment or evil is negative. It is to this class of concepts that liberty belongs: it describes the absence of a particular obstacle – coercion by men. It does not assure us of any particular opportunities, but leave it to us to decide what use we shall make of the circumstance in which we find ourselves.

Liberty as described by Hayek above represents its pure form. We may never escape coercion, but we can constrain it so as to maximize individual freedom and creativity:

Coercion is evil precisely because it thus eliminates an individuals as a thinking and valuing person and makes him a bare tool in the achievement of the ends of another… Coercion, however, cannot be altogether avoided because the only way to prevent it is by the threat of coercion. Free society has met this problem by conferring the monopoly of coercion on the state and by attempting to limit this power of the state to instances where it is required to prevent coercion by private persons. This is possibly only by the state’s  protecting known private spheres of the individuals against interference by others and delimiting these private spheres, not by specific assignation, but by creating conditions under which the individual can determine his own sphere by relying on rules which tell him what the government will do in different types of situations.

Coercion by some prevents individuals from enacting their will on reality. The advantage of rules that make predictable the employment of violence is that, although they cannot provide an individual with absolute freedom – what can? – rules greatly expand one’s option set.

Michel Foucault takes a different approach. As with Hayek, Foucault understands that institutions transform and delimit violence. But for Foucault, this transformation is merely in extension of state of war. Institutions embody oppression:

The role of political power, on this hypothesis, is perpetually to re-inscribe this relation through a form of unspoken warfare; to re-inscribe it in social institutions, in economic inequalities, in language, in the bodies themselves of each and everyone [sic] of us.
So this would be the first meaning to assign the inversion of Clausewitz’s aphorism that war is politics continued by other means. (Power/Knowledge: Selected Interview and Other Writings, 1972-1977, Two Lectures)

Foucault’s perspective is not unfounded. Recent research on violence and institutions reinforces his view. In their theory of development, North, Weingast, and Wallace state explicitly that the earliest institutions are those that extract rents for elites that are greater than can be extracted under a state of warfare:

To be credible, the commitment [to peace] requires that the violence specialists be able to mobilize and gather their rents, which are produced by the remainder of the population. Mobilizing rents, in turn, requires specialists in other activities. It is here that we move away from the simple ideas about violence and back toward a more reasonable depiction of the logic of the natural state. In the earliest societies of recorded human history, priests and politicians provided the redistributive network capable of mobilizing output and redistributing it between elites and non-elites. (Violence and Social Orders)

If this was the whole story, Foucault would be correct to claim that:

The system of right, the domain of the law, are permanent agents of these relations of domination, these polymorphous techniques of subjugation. Right should be viewed, I believe, not in terms of a legitimacy to be viewed, but in terms of the methods of subjugation that it instigates. (Power/Knowledge: Selected Interview and Other Writings, 1972-1977, Two Lectures)

The extremeness of Foucault’s claim here betrays him. “The system of right, the domain of the law” are not “permanent agents” but constantly evolving compromises whose benefits flow outward from  the elites who initiate them to broader sections of the population. Violence, in particular violence from elite groups, is the starting point. Systematization of violence may represent a more benign form of warfare removed one degree from violence, but as the agreements evolve over time, extractive rents are diminished and individual freedom can, though it is not necessary that it does, increase.

Foucault leaves the reader with a solution that is inefficient, if not altogether unworkable. For Foucault, the power of violence continues in the disciplining of society, and therefore “it is not toward the ancient right of sovereignty that one should turn, but towards the possibility of a new form of right, on which must indeed by anti-disciplinarian, but at the same time liberated from the principle of sovereignty.” Foucault asks for a quantum leap in social organization, an elimination of violence. His singular concentration on institutions as an extension of oppression disallows his recognition of gains made by broader portions of the population that have resulted under existing regimes. As institutions evolve, violence is constrained by rules and society may move closer to a state, as phrased by Hayek, “in which coercion of some by others is reduced as much as is possible in society.” That is a state of freedom. On the other hand, one ought to remain conscious of Foucault’s critique as it carries with it a public choice angle. Rules and laws often benefit particular parties to the disadvantage of others.

Thoughts on Matt Zwolinski's Argument for Basic Income Guarantee

The blogosphere is abuzz with critique of Matt Zwolinski’s welfare proposal. Matt suggests that the government guarantee a minimum income in place of the welfare state. He gives 3 reasons to support this proposal: 

1) A Basic Income Guarantee [BIG] would be much better than the current welfare state.  

2) A Basic Income Guarantee might be required on libertarian grounds as reparation for past injustice.  

3) A Basic Income Guarantee Might be required to meet the basic needs of the poor.

On the first one, the answer is “no duh.” Economists agree that straightforward cash transfers are more efficient than in-kind transfers. He captures the spirit of his argument in his final remark on the issue:

Shouldn’t we trust poor people to know what they need better than the federal government?

His second argument, that “a Basic Income Guarantee might be required” is not so obvious to me nor is this obvious to David Henderson:

I think we can all agree that many people have what they have at least in part due to previous rights violations. It doesn't seem clear to me that they are on top in what I take to be Matt's narrower sense rather than my wider sense. I think, for example, of people who paid into Medicare and Social Security only a fraction, even in present value terms, of what they get back from taxes on the current young and middle-aged people. Sure, many of them are on top, but many are not. I don't see how a basic income guarantee redresses that rights violation.

Finally, the third argument seems too strong of a statement. That “a Basic Income Guarantee might be required to meet the basic needs of the poor” is at best ambiguous. I am skeptical that a BIG is required for any purpose in regard to basic needs. If implemented, it will reflect that individuals prefer to live in a society with a basic safety net. It can be promoted without being wrapped in a moral argument.

We need to ask some of the most important question in political economy. Under what rules shall we live? How shall we decide these rules? The BIG, along with its funding, is a potential rule. If some threshold of society, say 2/3s, votes directly or indirectly for a BIG, than a BIG has been legitimized in some respect. (i.e., via constitutional amendment) Philosophical foundation is helpful, but I am not sure there is any better reason than that enough of the U.S. electorate has agreed upon this arrangement.

Why would we want a BIG? Consider the impact on two accounts. Matt Zwolinski hits on the first. If we are to have a safety net, direct transfers are preferable. As suggested by Hayek, any payment must be distributed equitably, although this obviously does not require equitable taxation – in such a case there would be no possibility of a BIG. If we have to choose, BIG is better. Second, there is reason to prefer a society that encourages individuals to take risks and innovate. To encourage individuals not to fear failure! A large minimum income might incentivize this too much, but a small guaranteed income, say $5000, would probably not represent an economically destabilizing moral hazard. It would represent relative subsistence, which is probably optimal for a welfare program.

Matt Zwolinski also notes problems with a BIG.

1) Disincentives [to work]  

2) Effects on Migration  

3) Effects on Economic Growth

A $5000 BIG would probably not entail problems with 1) and 3). Problem 2), that a BIG incentivizes migration, can be fixed by a simple rule. If you are an immigrant, you must wait 5 years before receiving the BIG.

Those who want to see a complete disassembly of government will have problems with all of the above. While I sympathize, I do not expect the government to go away any time soon. I do not expect welfare programs to disappear any time soon. So we have a choice. Libertarians can present realistic reform programs or simply allow politics to continue as usual as our ideas are ignored. 

A program like BIG forgoes political paternalism and the narrative of equality that has dominated the politics of the last century. It instead, promotes a political program that embraces individual autonomy.

***Late Thought: We could also define BIG as some percent of average per capita GDP. 5000 would be about 10% by today's estimates.

Hawtrey's Narrative and My Critique of the Classical Gold Standard/Monometallism

I’ve been making my way through one of Ralph Hawtrey’s classic works, The Gold Standard in Theory and Practice. The book is both highly readable and insightful. Of interest today is Hawtrey’s narrative of the classical gold standard.

Gold and silver were both used for exchange in commerce throughout much of recorded history. It was not until the 1870s that the western world moved to a uniform gold standard. The foundations for this move were laid as a result of large gold discoveries in Australia and California during the middle of the 19th century. Gold had been the more highly valued money, which under bimetallism meant that it was hoarded and silver was used for exchange as official par overvalued it. After the gold discoveries, the reverse was true. The increase in the monetary gold stock brought down its value so that now silver traded at a premium under bimetal standards. Hawtrey notes that this was not without consequence:

The French franc from a silver unit became a gold unit. The two great financial centres of the world, London and Paris, were both gold centres.

At the time Germany employed a silver standard, and conditions were not favorable for such a position:

Germany no longer derived any advantage from the silver standard in her trade with Eastern Europe, because silver had there made way for inconvertible paper. The bimetallic currencies of Western Europe had passed from a state of fixity in terms of silver [i.e., a de facto silver standard under a bimetal regime], with a slight fluctuating premium on gold, to fixity in terms of gold, with a slight fluctuating premium on silver… Even if the ratio of gold to silver continued to be stabilized by the bimetallism of the Latin Union, the silver standard might be expected to be a disadvantage to the German financial centres.

After its war in 1871 Germany adopted the gold standard and suspended the free coinage of silver. With new demand for gold, this shifted the price ratio in favor of a de facto silver standard for bimetal regimes. [See my brief discussion of Gresham’s law from last post.] That is, had bimetal regimes operated as they had beforehand, silver would have been employed under those regimes. But at this time, “Russia Austria-Hungary, Italy and the United States were using depreciated paper” and France made her money inconvertible. With the suspension of redemption in traditionally bimetal regimes:

There was nothing to relieve the sudden scarcity of gold, and the price of silver in gold began to fall. But that meant that the currencies of silver-using countries began to depreciate… The only remedy was the suspension of coinage of silver.

Given the above described political situation, the general adoption of the gold standard appears to be an accident of history. What were the precise impacts of this "accident"?

An interesting, yet not broadly acknowledged, aspect of the adoption of the gold standard concerns its promotion of a relatively unstable price level. After the gold standard was adopted, the impact on the price level from a change in the monetary gold stock or in production appears to increase. Notice the change that occurs in 1873:

Notice that prices had stabilized from about 1856 to 1873. The gold standard era can be divided into an era of price deflation, 1873-1896, and of price inflation, 1896-1914. During 1914 most nations on the gold standard suspended redemption, but since gold was implicitly the only metallic standard - the return of redemption was anticipated after the war - price rose steeply as demand for gold fell. Precisely what one would expect given a lack of metallic money substitutes. 

I ran a series of regressions to investigate:

Years	VARIABLES	British GDP	Monetary Gold Stock	Constant
1839	Sauerbeck-Statist Index	-0.608*	0.469**	9.583***
to		(0.32)	(0.19)	(2.78)
1873	SIGMA2			0.00701***
1839	Sauerbeck-Statist Index	-1.433***	0.878***	17.04***
to		(0.17)	(0.12)	(1.40)
1896	SIGMA2			0.00800***
1839	Sauerbeck-Statist Index	-1.404***	0.868***	16.76***
to		(0.14)	(0.10)	(1.16)
1913	SIGMA2			0.00718***
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

I used an ARCH model to account for the clustering of price level volatility as changes in direction tend to occur for more than one year at a time. The coefficients (those numbers with *s next to them) estimate the percent impact of a one percent increase in the independent variables – GDP and Monetary Gold – on the Sauerbeck-Statist Index, a measure of the British wholesale prices. The effect of both about doubles across the entire time period after the gold standard is established. The effects decrease slightly when the time period is extended to 1913, likely due to the increase in production of gold that began to occur in the 1890s after the discovery of the cyanide process for extraction. This alleviated the impacts of an increase in demand for gold.

Bottom line, a monometallic standard is bad for price stability. Was this bad for production? It is hard to tell because the industrial revolution was in full bloom at the time of the change. Clearly the world economy was able to handle a less flexible monetary standard as the growth associated with the period was before unthinkable. And perhaps the move to a common standard decreased transactions costs enough to compensate for the inflexibility. The interesting question is: on net, did the gold standard hurt or help economic growth?

Late note: The use of ARCH does not seem to matter as I receive the same output with a simple regression with Newey-West HAC errors.