Chapter 25: Transmission Mechanisms of Monetary Policy: The Evidence

Fluctuations in output and prices are undesirable for an economy and one goal of monetary policy is to minimize these fluctuations as economic stability promotes long-run economic growth. In order to be successful, economists need to understand exactly how changes in the money supply effect prices and output, known as the transmission mechanisms of monetary policy. In this chapter we evaluate the empirical evidence about the links between money and the economy along with many of the alternative transmission mechanisms.

I.  Framework for Evaluating Empirical Evidence

There are two types of empirical evidence about the link between money and the economy. This evidence heightens the debate over the impact of monetary policy because supporters of one type of evidence will not recognize the validity of other types of evidence.

Structural Models

Most Keynesians favor structural models to examine the effect of money on the economy. Structural models build a model of the economy to understand exactly how monetary policy effects the economy. For example, changes in the money supply change the equilibrium interest rate, which in turn changes investment spending and output:
iIY
Note how the structural model lays out exactly how money affects output: through its effect on interest rates and investment.

Reduced-Form Models

Monetarists tend to favor reduced-form models. Reduced-form models look directly for a relationship between money and output, but do not attempt to explain how this relationship develops. The "how" of this issue is simply left unanswered, sort of a mysterious "black box":
M?Y
So we know money is linked to output, but we do not examine why in a reduced-form model.

Advantages/Disadvantages

So which model is better? There are advantages and disadvantages to both approaches. The advantages of structural models over reduced-form evidence come from specifying the transmission mechanisms. By doing so structural models

However, structural models of one big disadvantage: Structural models only work if they are properly specified. That is, if they correctly identify all of the transmission mechanisms of monetary policy. If not they will come to inaccurate conclusions about the impact of money on the economy. That is a pretty big "if."

Reduced-form models do not have this specification problem since they are looking directly at the relationship between money and output and not how it occurs. If transmission mechanisms are numerous, complex, and constantly changing this would make structural modeling almost impossible, but it would not affect reduced-form modeling. This is the main advantage of reduced-form models.

But reduced-form models have their own problem: While reduced-form evidence shows a correlation between money and output, it does not necessarily prove causation. If money and output are show to move together, do increases in the money supply cause output to rise OR does rising output push up the money supply? Also, it is possible that a third variable affects both money and output, causing them to move together. Reduced-form evidence cannot always distinguish between these alternative scenarios.

II.  Early Keynesian Evidence

Early Keynesians (in the 1950s and 60s) came to the conclusion that money did not affect output. That is, monetary policy does not affect the business cycle. This belief was based on several pieces of evidence:

There are some problems with these conclusions. First, lets consider the Great Depression evidence: In terms of the relationship between nominal interest rates and investment. Monetarists would argue that the real interest rate, since it is the real cost of borrowing, is a much more important factor in investment decisions.

Finally, early Keynesian evidence only allows for one way for monetary policy to affect output. There may be other transmission mechanisms.

III.  Early Monetarist Evidence

In the 1960s, Friedman and other monetarist published evidence contradicting the popular Keynesian view that monetary policy did not matter. Your textbook summarizes their evidence in 3 different categories:

Timing

This evidence looks at whether movements in output followed movements in the money supply. Friedman and Schwartz found in 1963 that in EVERY recession, money supply fell 3-20 months before output did. They concluded that changes in money supply caused output to change, although the lag time varied.

The problem with this evidence is that, again, a third factor may be causing both money and output to change. The change in money supply is not necessarily exogenous, or independent of other factors in the economy. It could be that those other factors trigger the money-output relationship. The sequence of events could be in the reverse: money supply may be following previous output changes.

Statistical

This evidence looks at the correlation between money supply and output versus other factors and output to gauge the relative importance of money supply. Again Friedman did some of the research in the 1960s, and concluded that the money supply was highly correlated with the money supply, while Keynesian spending variables (investment and government spending) were not. Again, this study was subject to some criticisms about the chain of causation.

Historical

This evidence is probably the most compelling of all the early monetarist evidence. Friedman and Schwartz single out historical episodes where the change in the money supply was likely exogenous. (The 1936-37 changes in the reserve requirement is one such example) They still find that recessions follow exogenous decreases in the money supply.

Despite the criticisms of the monetarist evidence, the research of Friedman and others forced Keynesians to modify their views and acknowledge that monetary policy did indeed matter for business cycles.

IV.  Transmission Mechanism of Monetary Policy

Later research on structural models worked on developing addition channels through which money supply impacts output. Understanding all of the possible channels gives policy makers a best estimate of the full impact of monetary policy and how those impacts will change over time.

Interest Rate Channels

The modified Keynesian interest rate channel would look like this:
M increases ir decreases I increases Y increases
where ir is the real interest rate. A lower real interest rate would also encourage spending on consumer durable goods (like cars and appliances). The increase in money supply leads to higher expected inflation and thus lower real interest rates, even if nominal interest rates do not budge:
M increases exp. inflation increases ir decreases I increases Y increases
The importance of the interest rate channel is still a source of disagreement among economists. However, one of its biggest believers is John Taylor, the likely frontrunner to replace Alan Greenspan when he steps down as chair of the Federal Reserve.

Asset Price Channels

Changes in the money supply may also affect other asset prices (by change interest rates, changes in the money affect short-term bond prices).

If an increase in money supply lowers the real interest rate, the exchange rate will depreciate (see chapter 7). This makes imports more expensive here and U.S. exports cheaper abroad. The result is a rise in net exports and output:
M increases ir decreases dollar decreases net exports increase Y increases

Changes in the money supply could also affect stock prices. Tobin's q theory develops a link between money and output via a change in stock prices. Tobin's q is the ratio of a firms market value to the cost of new capital. If q is high, a firm will issue stock to buy new capital, and investment will rise. If q is low a firm will buy up older capital of existing firms and investment will fall. An increase in the money supply will increase spending, including spending on stock. Higher demand for stock increases stock prices and q. A higher q increases investment and output:
M increases stock prices increase q increases I increases Y increases

Higher stock prices also increase consumer wealth, which increases consumption and output:
M increases stock prices increase wealth increases C increases Y increases

Empirical research has supported the importance of exchange rate and wealth effects.

Credit Channels

Asymmetric information in the credit markets and the role of banks in reducing asymmetric information problems creates another set of channels through which changes in the money supply may affect output.

Increases in the money supply increase bank reserves and deposits, making more funds available for lending. Firms will find it easier to obtain loans for investment projects:
M increases bank deposits increasebank loans increase I increases Y increases
Through this channel, changes in the money supply will especially affect the investment decisions of small firms, who are more likely to be dependent on bank loans. Large firms raise capital directly through the stock and bond markets.

Increases in the money supply, by increasing stock prices, increase the net worth of firms. Higher net worth means banks have higher collateral, reducing adverse selection and moral hazard problems so banks will be more willing to lend. Increases in the money supply also increase a firm's cash flow through the reduction in short-term nominal interest rates. Higher cash flow also reduces adverse selection and moral hazard problems so banks will be more willing to lend:
M increases stock prices increase and cash flow increases adverse selection & moral hazard decrease
lending increasesI increases Y increases

For households, there is also a balance sheet affect. Higher stock prices improve the household balance sheet, making financial distress less likely. Given this, households will be willing to hold more illiquid assets like consumer durables (cars, appliances, houses) and fewer liquid financial assets. The purchase of durables will increase output:
M increases stock prices increase probability of financial distress decreases
C increases Y increases

All of these channels are summarized in figure 3, page 351.
Wow, that's a whole lot of channels! Money supply affects output through nominal and real interest rates, stock prices, exchange rates, and possibly other ways. It is now easy to see why modeling the effects of monetary policy is a very complex job.

Lessons for Monetary Policy

Despite the different types of conflicting evidence, research on the transmission mechanisms of monetary policy does yield several points of agreement among economists:

FYI: Related Links

Financial Innovation and Monetary Transmission The site contains papers from a conference at the FRBNY. Obviously, the Fed must keep up with changes that affect how monetary policy works.

The Wealth Effect Take a closer look at the phenomenom that helped drive the economic expansion of the 1990s.