Scott Sumner  

David Beckworth interviews Steve Horwitz

PRINT
Walter Block's Reductio ad Abs... Friedman in 1946...

David Beckworth's podcast series continues to produce fascinating interviews. Today I'll discuss a recent interview with Steve Horwitz. Many of Steve's views coincide with market monetarism, although in a few areas his views lean more in the Austrian direction. Here are some things I noticed:

1. Steve and David talked about how "money touches everything", which means it is involved in almost all transactions. That's what makes money special. This is a point Nick Rowe also emphasizes. I think that's right, although I focus more on the fact that goods and services are priced in money terms, and a bit less on the role of money as a medium of exchange. (The two approaches are hard to disentangle.)

2. There was a lot of discussion about the problem of monetary disequilibrium. A root cause of this is short run price stickiness, which prevents the economic system from smoothly adapting to a sudden change in money supply, or (equally important) money demand. Steve rightly emphasized that this is a "coordination problem". Again, I think this is basically right, but one word of caution. Monetary disequilibrium is a slippery concept, very different from disequilibrium in the housing market when there are rent controls. When there is a sudden increase in the money supply, you might say that people hold more money than they wish to. Then they try to get rid of these excess balances through spending, which drives up aggregate demand.

That's a story I also use. But how do you identify monetary disequilibrium? At times, Steve talked in terms of the price level being the right indicator. But then later he suggested (correctly in my view) that deflation might not be a problem if driven by strong productivity growth that lowers the cost of production. That leads to NGDP perhaps being the right metric. Steve switched over to the NGDP approach when he started using the equation of exchange as a framework, and M*V as the monetary indicator. So what is the right indictor of monetary disequilibrium? An unstable price level, an unstable NGDP, or something else?

I look at the disequilibrium process somewhat differently. I see the labor market as being the area most strongly impacted by monetary shocks. Thus a decline in the money supply doesn't make it hard for people to get the cash holdings they prefer (interest rates will rise until there is enough cash in ATMs for anyone who wants it) but it will make it hard for the unemployed to find the jobs they want, and indeed could find in times when monetary policy is more stable. That's the real disequilibrium problem.

3. Steve complained that the Keynesian tendency to draw money S&D diagrams with interest rates on the vertical axis (instead of 1/P) obscures the role of money as a medium of exchange. That's my view as well. Unfortunately our textbooks increasingly emphasize interest rates and de-emphasize the money supply.

4. There was a discussion regarding the question of what is the correct money supply. Steve suggested that, at a minimum, it ought to include assets that are media of exchange, such as demand deposits. He also discussed recent ideas such as divisia indices, which are weighted averages of each type of money.

I don't think debates over the question, "What is money?" are very useful. (That which has no practical implications, has no theoretical implications.) What matters are very specific questions, such as what sort of monetary policy provides macroeconomic stability. The answer to this policy question in no way hinges on whether the base or M1 or M2 or MZM are closer to the platonic ideal of "money". The Fed directly controls the base, and indirectly influences M1, M2 and NGDP. The relevant question is what sort of policy for controlling the base results in the best macroeconomic outcome. The answer is not at all likely to include "transactions money", such as M1 or M2.

5. I probably put less weight on Cantillon effects than Steve does, but let me agree on a few points. I do think that monetary shocks have relative price effects, if only because they have cyclical effects (impacting RGDP). That's true regardless of how the money is introduced into the economy. Second, there may be monetary regimes where Cantillon effects matter a lot, especially where the central banks goes far beyond the traditional bounds of purchasing Treasuries.

6. I am less worried than Steve about bad monetary policy fooling people into malinvestment. In my view the public knows much more than the Fed. They saw the Fed was off course in 2008 well before the Fed itself understood the problem. So we don't live in a world where central banks fool the public. Yes, the public does occasionally make bad investments, but these cannot be predicted by outsiders, even outsiders who notice that monetary policy is off course. The public knows that too!

In other words, any sensible economist or investor knew that monetary policy in 1968-69 was too expansionary, but this knowledge was useless when contemplating which stocks to short of Wall Street. Malinvestment is really hard to spot, because markets are very, very efficient. I suppose this is one area where I disagree with many Austrians.

7. Steve also discussed "endogenous busts", which means something like "the boom creates the bust." I have more sympathy for this view than many Keynesians, but in the end my position is probably better described as intermediate between the Keynesian and Austrian positions. Keynesians see the economy as a vast featureless plain with occasional canyons created by spending shortfalls. Austrians focus on how the preceding boom created the recession, at least in many cases---like mountains rising above a plain. I worry when I see things like the following (from a recent edition of The Economist):

The only way to know if America can manage a repeat performance is to test the economy's limits. The transition from a 2% target to a higher one would offer a chance for such an experiment. As it is, a central bank hell-bent on keeping inflation low and stable risks cutting short a boom with room to run.
Yes, money was too tight in 2008, and for a long period afterwards. But let's not forget the lessons of the 1960s and 1970s. This sort of experiment is not without risk. If we overshoot we should not expect a smooth landing on a higher upland plateau, rather a sharp correction into recession. If we are going to do this type of experiment, I'd rather start from a position of more than 4.4% unemployment. Given where we are today, steady as you go seems the most prudent course.

At the risk of being unfair to both groups, I sometimes feel that the Keynesians think recessions are caused by valleys, and Austrians think recessions are caused by mountains. I think recessions are caused by uneven ground. If NGDP is more than expected at one point in time, then there will be other times when it is less than expected. And those will be periods of high unemployment. Thus a sudden increase in NGDP can be just as destabilizing to an economy as a sudden decease. Sometimes an Austrian economist will go too far with this---blaming too much of the Great Depression or Great Recession on the preceding boom, but it's a valid concern.

Keynesian view: We were doing fine, until we fell into a deep canyon:

Screen Shot 2017-07-10 at 6.18.54 PM.png
Austrian View: Because we walked up the left side of Kilimanjaro, we then had to slide down the right side:

Screen Shot 2017-07-10 at 6.18.07 PM.png
My view? Avoid mountains and valleys. Bonneville Salt Flats is your goal.

PS. I've relied on my memory and a few notes--apologies if I misrepresented Steve's views.


Comments and Sharing






COMMENTS (10 to date)
Steve Horwitz writes:

No apologies necessary, Scott. Your memory and notes were dandy - I think you have represented me accurately. Glad you enjoyed it!

Scott Sumner writes:

Thanks Steve. It was a very nice talk--ideas very clearly explained.

dlr writes:

Great post Scott.

Andrew_FL writes:
That which has no practical implications, has no theoretical implications.

This is an extremely puzzling statement. It is not the purpose of theory to limit itself to what is relevant to the peculiar circumstances of time and place.

Steve switched over to the NGDP approach when he started using the equation of exchange as a framework, and M*V as the monetary indicator.

Not really, the domestic currency denominated gross nominal value of purchases on newly produced final goods is not obviously the most natural measure of the Effective Money Stream.

To be honest it seems exceedingly unlikely to me that which transactions should "count" happens to coincide exactly with what we happen to have the longest measure of.

I am less worried than Steve about bad monetary policy fooling people into malinvestment. In my view the public knows much more than the Fed. They saw the Fed was off course in 2008 well before the Fed itself understood the problem. So we don't live in a world where central banks fool the public. Yes, the public does occasionally make bad investments, but these cannot be predicted by outsiders, even outsiders who notice that monetary policy is off course. The public knows that too!

Malinvestment happens because of knowledge neither the public nor the central bank possess. The central bank's policy destroys information.

In other words, any sensible economist or investor knew that monetary policy in 1968-69 was too expansionary, but this knowledge was useless when contemplating which stocks to short of Wall Street. Malinvestment is really hard to spot, because markets are very, very efficient. I suppose this is one area where I disagree with many Austrians.

Malinvestment is difficult to spot but that has nothing to do with markets being "very, very efficient"-it's because by the time the relevant information is discovered, misallocation of the misallocation of real resources into the wrong physical production goods has already occurred.

Of course it's impossible to understand that working implicitly from a Knightian capital theory. Although I'm not sure if that's the problem here so much as that you think of investment purely in terms of the stock market for some reason?

How far along the structure of production is a given company's stock? AFAICT this question doesn't make sense.

I don't like the topography analogies but it's difficult for me to articulate why.

Thaomas writes:

Why would deflation resulting from a positive technological or terms of trade (eg China in the WTO) shock not be something that a central bank ought to offset? If uncertainty about future price levels have costs (difficulties, given sticky prices, of adjusting to relative price changes), why not prevent the uncertainty regardless of its source?

As for whether current Fed policy is correct, would this not be a good time to show that they really do have a PL target and not an inflation rate ceiling?

denis writes:

[Comment removed. Please consult our comment policies and check your email for explanation.--Econlib Ed.]

Scott Sumner writes:

Thanks dlr

Andrew,

I am including any conceivable practical implications, even if not measurable with current technology. Thus does "string theory" have practical implications that might be measured with some conceivable future technology? I don't know--but that would be the test.

I don't follow your point about NGDP---I was just describing Steve's view. (Although I happen to agree.)

If the knowledge is destroyed, and no one has it, how can we confidently assert that an excessively expansionary monetary policy led to an unsustainable boom, followed by a bust?

I don't think of investment purely in terms of the stock market, it's just that stocks provide a convenient way to bet on the "malinvestment" hypothesis.

Thaomas, Because that sort of deflation does not affect NGDP, which is what really matters for labor and financial market stability.

Andrew_FL writes:
I don't follow your point about NGDP---I was just describing Steve's view.

I could be wrong, and I may have to listen to the podcast again, but I don't believe he actually says "GDP" or "NGDP" in this context at any point.

My point is that in constructing an exchange equation and selecting a measure or proxy for the effective money stream, anything less than choosing the Paleo-Monetarist "PT"-all transactions denominated in the relevant currency-one necessarily selects some transactions as for some reason "counting" as opposed to others. It is not obvious why monetary disequilibrium theory would be a theory about money transactions on final new domestically produced goods only.

If the knowledge is destroyed, and no one has it, how can we confidently assert that an excessively expansionary monetary policy led to an unsustainable boom, followed by a bust?

The information about the correct array of nominal prices necessary to coordinate market activity is destroyed by monetary policy and the "new" correct set of prices must be found by a process of entrepreneurial discovery. It is only *after* entrepreneurs have discovered this new information that we can retrospectively say an unsustainable boom occurred.

We can observe that certain typical patterns repeat themselves in historical business cycles-I think the evidence is very strong that the overall structure of the production process broadly conceived is significantly distorted both on the up and down swing, in pretty much the way theory would expect.

I don't think of investment purely in terms of the stock market, it's just that stocks provide a convenient way to bet on the "malinvestment" hypothesis.

I don't think they do. Again, I don't know that one can look at a stock price for an individual company and say "this company as a whole is at this point on the structure of production" or "is [x] degrees removed from final consumption." Most firms are more diversified than that.

Scott Sumner writes:

Andrew, I think he said "MV" which implies NGDP. And see the first comment above.

Andrew_FL writes:

@Scott Sumner-You are surely aware that in the history of the equation of exchange, setting MV = NGDP is a relatively recent development. Fisher's PT was much broader, for example.

POST A COMMENT




Return to top