Unconventional Monetary Policy, circa 1793

From Henry Thornton’s An Inquiry Into The Nature And Effects Of The Paper Credit Of Great Britain:

The truth of this observation, as applied to Bank of England notes, as well as the importance of attending to it, may be made manifest by adverting to the events of the year 1793, when, through the failure of many country banks, much general distrust took place. The alarm, the first material one of the kind which had for a long time happened, was extremely great. It does not appear that the Bank of England notes, at that time in circulation, were fewer than usual. It is certain, however, that the existing number became, at the period of apprehension, insufficient for giving punctuality to the payments of the metropolis; and it is not to be doubted, that the insufficiency must have arisen, in some measure, from that slowness in the circulation of notes, naturally attending an alarm, which has been just described. Every one fearing lest he should not have his notes ready when the day of payment should come, would endeavour to provide himself with them somewhat beforehand. A few merchants, from a natural though hurtful timidity, would keep in their own hands some of those notes, which, in other times, they would have lodged with their bankers; and the effect would be, to cause the same quantity of bank paper to transact fewer payments, or, in other words, to lessen the rapidity of the circulation of notes on the whole, and thus to encrease the number of notes wanted. Probably, also, some Bank of England paper would be used as a substitute for country bank notes suppressed.

The success of the remedy which the parliament administered, denotes what was the nature of the evil. A loan of exchequer bills was directed to be made to as many mercantile persons, giving proper security, as should apply. It is a fact, worthy of serious attention, that the failures abated greatly, and mercantile credit began to be restored, not at the period when the exchequer bills were actually delivered, but at a time antecedent to that æra. It also deserves notice, that though the failures had originated in an extraordinary demand for guineas, it was not any supply of gold which effected the cure. That fear of not being able to obtain guineas, which arose in the country, led, in its consequences, to an extraordinary demand for bank notes in London; and the want of bank notes in London became, after a time, the chief evil. The very expectation of a supply of exchequer bills, that is, of a supply of an article which almost any trader might obtain, and which it was known that he might then sell, and thus turn into bank notes, and after turning into bank notes might also convert into guineas, created an idea of general solvency. This expectation cured, in the first instance, the distress of London, and it then lessened the demand for guineas in the country, through that punctuality in effecting the London payments which it produced, and the universal confidence which it thus inspired. The sum permitted by parliament to be advanced in exchequer bills was five millions, of which not one half was taken. Of the sum taken, no part was lost. On the contrary, the small compensation, or extra interest, which was paid to government for lending its credit (for it was mere credit, and not either money or bank notes that the government advanced), amounted to something more than was necessary to defray the charges, and a small balance of profit accrued to the public. For this seasonable interference, a measure at first not well understood and opposed at the time, chiefly on the ground of constitutional jealousy, the mercantile as well as the manufacturing interests of the country were certainly much indebted to the parliament, and to the government.

What is the Mechanism?

What is the mechanism by which monetary policy affects nominal and real variables? Given the recent debates over monetary policy, it is clear that not only is there no consensus on the issue, but there is actually widespread divergence among economists. However, the issue of the monetary transmission mechanism is central to understand how, if at all, monetary policy can be effective in achieving its intended goals.

The mainstream, New Keynesian view suggests that monetary policy works through its effect on the real interest rate. In normal times, the central bank adjusts the nominal interest rate. Since prices are sticky, this has a corresponding effect on the real interest rate. Since the real interest is negatively correlated with consumption and investment, this has real effects in the short run. In the long run, inflation expectations adjust and the effect of policy is purely nominal.

According to this view of the monetary transmission mechanism, monetary policy becomes impotent (or at least more difficult) at the zero lower bound on the nominal interest rate. When the interest rate is at zero, the only way that policy can be effective in this context is by influencing inflation expectations because an increase in short-run inflation expectations (holding the nominal interest rate constant) can also reduce the real interest rate.

This is the logic that underscores many calls for targeting the price level or the level of nominal GDP. The appeal of such policies, according to this view, is that they raise short-run inflation expectations while keeping long-run expectations anchored. Thus, one can generate short-run real effects without sacrificing long-run objectives of price stability.

This raises two main questions. First, how can monetary policy achieve this goal? Second, what evidence do we have to suggest that this type of policy would be successful?

According to the New Keynesian view, the ability of monetary policy to affect inflation expectations is captured in the central bank’s credibility. If the central bank announces a price level or nominal GDP target, the policy is only effective in the NK case if the public believes that the central bank is committed to that target. According to some, all this requires is an announcement by the central bank. If the central bank lacks credibility, the only way to convince the public to expect higher short run inflation is to engage in a policy consistent with the price level or nominal GDP target. However, in the absence of credibility, it is unclear how this happens in the NK model. In other words, in the NK model the nominal interest rate is stuck at zero and therefore the only way that monetary policy can have any effect is through influencing inflation expectations. If the central bank lacks credibility, it cannot raise such expectations. Also, given the monetary transmission mechanism in the NK model, monetary policy is not even capable of convincing the public that it is serious about its target because there is no action that it can take to achieve its goal. This thereby reinforces the public’s view about the credibility of the central bank. This is a liquidity trap.

But what if the NK model of the monetary transmission mechanism is wrong? Suppose that we have another view of the monetary transmission mechanism. Under what conditions can monetary policy be effective and what role would inflation expectations play? My previous post was an attempt to motivate this discussion. In that post I referenced a working paper of mine that examines the role of open market operations in influencing the price level and nominal spending. In that paper, monetary policy is transmitted through the aggregate stock of liquid assets (i.e. assets used in transactions). Thus, the ability of monetary policy to affect the price level or nominal spending is dependent upon whether open market operations have an effect on the aggregate supply of liquidity. In other words, if money and bonds are perfect substitutes, then open market operations are ineffective. If, however, money and bonds have differing degrees of liquidity, then open market operations are always effective (i.e. the effects are not dependent on inflation expectations and the zero lower bound is irrelevant). Thus, determining whether monetary policy can be effective has nothing to do with expectations or the nominal interest rate, but only whether money and bonds are perfect substitutes. In addition, regardless of whether money and bonds are perfect substitutes, it is possible for open market operations to be effective if the central bank buys other, less liquid assets.

Clearly, the view put forth in my working paper differs from the monetary transmission mechanism in the NK model. This brings us to the second question above as well as a follow-up question. What evidence do we have to suggest that adopting a price level or nominal GDP level target would be successful? How can we understand this evidence in light of the NK model and the view put forth in my working paper?

The most frequently cited example is that of FDR raising the price of gold in 1933. To understand the effects of this change, it is useful to revisit my post, “Commodity Money: A Primer.” The workings of the gold standard can be described as follows (taken from that post):

Suppose that resources can be used to produce consumer goods and gold.  For a given amount of resources, R_fixed, we can write a production possibilities frontier:

R_fixed = R(C, G)

Where C is consumer goods and G is gold.  The marginal productivities can be of producing consumer goods and gold are, respectively, R_c > 0 and R_g > 0.

Resource owners earn income:

Y = (P_c/P_g) C + G

Where Y is income and P­_c and P_g are the prices of consumer goods and gold, respectively.

Equilibrium necessarily requires that

R_c/P_c = R_g/P_g

Or

R_c/R_g = P_c/P_g

This implies that the price of consumer goods in terms of gold is equal to the marginal opportunity cost (marginal rate of transformation).

Given this equilibrium condition, if the price of gold is exogenously increased, the price level will have to rise. But what is the mechanism by which the price level rises?

According to the New Keynesian view, this is a prime example of level targeting. The government announces a higher price of gold, which requires a higher price level. Short-run expectations of inflation rise. Long-run expectations about price stability are anchored. This is a textbook case of level targeting.

Nonetheless, it is not clear that this expectations channel is actually the correct transmission mechanism. For example, if monetary policy is transmitted through aggregate liquidity, then the increase in the price level results from the corresponding exogenous increase in the monetary base. Along these lines, the traditional Old Monetarist transmission mechanism would suggest that the real effects of the policy are directly the result of the increase in the monetary base because of portfolio/distributional effects on money balances.

Each of these transmission mechanisms provide (somewhat) observationally equivalent results. The precise mechanism through which monetary policy works is important given the different monetary regimes. For example, if monetary policy works through expectations, then the ability of current monetary policy to be successful is to influence inflation expectations. Under the gold standard, all this requires is a change in the price of gold. Under the current policy regime, this requires that central bank has sufficient credibility to convince the public that they are committed to some level target.

If monetary policy works through its effect on aggregate liquidity, under a gold standard the increase in the price of gold results in an increase in the monetary base. Under the current monetary regime, monetary policy must be conducted through open market operations. The ability of the central bank to have an effect on the aggregate stock of liquidity is therefore dependent on the relative liquidity properties of base money and the assets purchased in the open market.

What this tells us is that we cannot completely understand the events of the 1930s without understanding the monetary transmission mechanism. The evidence suggests that monetary policy can be effective (very effective). Nonetheless, the different in monetary regimes are important. Under the gold standard, the change in the price of gold is sufficient to produce an exogenous increase in the monetary base and short-run inflation expectations. Under the current monetary regime, influencing expectations requires credibility on the part of the central bank. In addition, increasing the aggregate stock of liquidity through open market operations depends on the relative liquidity property of money and bonds (and other assets).

In discussing policy and providing policy advice, it is first important to understand the monetary transmission process. Unfortunately, there is little agreement about how monetary policy is transmitted and existing evidence often supports mechanisms that are observationally equivalent, but yet very different processes. Perhaps this is why there are such divergent views on the current stance of monetary policy.

Can More Monetary Policy Be Effective?

Suppose that we accept two basic arguments:

1. The central bank’s objective is to target the price level or nominal income.

2. The price level/nominal spending is below the optimal time path.

[A quick note: This does not mean that one should accept these arguments. Rather, I want to consider whether expansionary monetary policy can be effective by first assuming that we would find such a policy desirable. This allows us to focus exclusively on the role of monetary policy effectiveness.]

These two points imply that the central bank has a stated objective and is short of that objective. Expansionary monetary policy is necessary. However, even if we accept these basic points, there are other questions that need to be answered:

a. What is the monetary transmission mechanism? According to the New Keynesian model, monetary policy is transmitted through the interest rate. Once the interest rate hits the zero lower bound, monetary policy is relatively ineffective. Recent rounds of quantitative easing have been sold in a similar respect with promises of lower long term interest rates. However, this type of explanation assumes that there is something that prevents arbitrage in the market for government debt (i.e. that people holding long term bonds are different people from those holding short term bonds). This seems a weak reed upon which to rest policy. In the absence of this assumption it is difficult to understand how this policy might be effective in achieving the goal of monetary policy.

An alternative view of the monetary transmission mechanism that the central bank influences economic behavior through the aggregate stock of liquidity. According to this view, it is the aggregate stock of liquidity that determines the price level/nominal GDP.

b. Taking the monetary transmission mechanism in point (a) as given, can open market operations affect the price level/nominal GDP? It would seem that (according to this view) open market operations could be effective if they affect the aggregate stock of liquidity. But under what conditions do open market operations affect the aggregate stock of liquidity?

This brings me to a working paper of mine that examines whether or not open market operations can be effective. Here is the abstract:

Recent rounds of quantitative easing by the Federal Reserve have sparked debate about the effectiveness of such policies. Whether or not such policies can be effective depends on the monetary transmission mechanism. The present paper considers the effectiveness of open market operations when monetary policy is transmitted through the aggregate stock of liquidity. It is shown that monetary policy is effective when the liquidity of money and bonds differs. If bonds and money share the same liquidity properties, however, open market operations are irrelevant. These results hold regardless of whether liquidity is exogenous or endogenous. More generally, monetary policy is most effective when the central bank buys relatively illiquid assets. Even if open market operations are irrelevant, there is an empirical link between the aggregate stock of liquidity and the price level and nominal spending. If money and bonds are perfect substitutes in terms of liquidity, expansionary fiscal policy can influence the price level and nominal spending through its effect on liquidity. When liquidity is endogenous, fiscal policy is irrelevant if money and bonds are imperfect substitutes.

In short, the model shows that open market operations are effective in influencing the price level and nominal spending when money and bonds are not perfect substitutes for liquidity. This is true regardless of whether liquidity is exogenous or endogenous. However, if money and bonds have the same properties in terms of liquidity, open market operations only affects the composition of the aggregate stock of liquidity. As I point out in the paper, it actually seems difficult to empirically test whether or not open market operations affect the aggregate stock of liquidity.

I also consider the implications for fiscal policy, economic welfare, quantitative easing, and also compare the results to similar models in the literature.

The paper is only preliminary, so any and all constructive comments on the paper would be welcome. Hopefully, this can also spur some debate in the blogosphere — in a positive direction — while simultaneously demonstrating the way I have been thinking about monetary policy and why it doesn’t fit with traditional New Keynesian-type analysis.