Category Archives: Mark Thoma is Indispensable

Money, Inflation, and Wile E. Coyote

Posted on by

Before 2008, the bubble that some economists expected to pop was the value of the dollar. Paul Krugman used a colorful metaphor to describe this.

So it seems likely that there will be a Wile E. Coyote moment when investors realize that the dollar’s value doesn’t make sense, and that value plunges.

I found this in an old post from Mark Thoma. Feel free to use Google to find other citations.

Nowadays, when he looks at the prices of U.S. government bonds, Krugman sees rational expectations at work. He looks at the low long-term interest rates as a sign of the market’s wisdom in predicting low inflation for ten years.

But what I see is a market that could have a Wile E. Coyote moment. Once enough investors decide to dump our bonds, interest rates will rise. It will become clear that at those interest rates the government cannot afford to pay off the bonds, so more investors will dump bonds. etc.

Now the Fed can always buy our bonds. It is doing a lot of that, and I can see where an investor with a sufficiently short time horizon who believes that he won’t be the one still holding bonds when they start to lose value might say, “Don’t fight the Fed. Just ride the yield curve for a little while longer.”

But suppose that our Wile E. Coyote moment comes when inflation has been heating up. (Indeed, the Wile E. Coyote moment could come because inflation heats up, and at that point investors decide that the Fed may no longer be their friend.) Under this scenario, the Fed wants to be a bond seller, not a bond buyer, in order to keep inflation in check. If the Fed feels constrained not to sell too many bonds, then inflation could really take off.

How can I justify a fear of inflation, given recent behavior? In recent years, the Fed has created a lot of money, and we have not seen a lot of inflation. What is going on?

1. Perhaps money and inflation have no connection. We should go back to using the Phillips Curve. When unemployment is high, inflation is low, and conversely. After all, wages are 70 percent of costs, and it seems unlikely that wage inflation will get much traction with folks having a hard time finding jobs.

2. Perhaps we are suffering from tight money. This is the Scott Sumner argument. Money and inflation (he would prefer nominal GDP) are related, inflation is low, ergo we must have tight money.

For the short run, I believe something like (1). However, I am old enough to remember the 1970s. Based on that experience, I would say that inflation is subject to regime shifts. There is a regime in which inflation is low and relatively stable. There is another regime in which inflation is high and volatile. Finally, there is a regime of hyperinflation.

I think that with enough persistence, the Fed can move us between the low, stable regime and the high, volatile regime. The Fed spent the 1970’s getting us into the high, volatile regime, and it spent the 1980s getting us out of it.

Hyperinflation is a fiscal phenomenon. A government that can balance its budget is never going to have hyperinflation.

The scenario I have in mind is one in which the economy has begun to shift to the regime of high and volatile inflation. Then the Wile E. Coyote moment arrives, and the Fed feels pressed to keep the U.S. bond market “orderly” by not selling bonds. In fact, interest rates are rising so quickly that the Fed decides that it needs to buy bonds. This sets off a spiral of money-printing and price increases, threatening to bring on hyperinflation. In which case, the bond market will not be orderly. Nor will anything else.

Fake Wealth

Posted on by

Michael Munger writes,

If you measure from the peak of the bubble, we lost a lot of wealth. But that wealth was entirely fake, created by a revved up demand for houses as assets expected to appreciate rapidly.

Pointer from Mark Thoma. As Tyler Cowen would say, “We’re not as wealthy as we thought we were.”

If you are a Keynesian, fake wealth is a good thing. It is well known that deficit spending is pretty much useless if there is “Ricardian equivalence,” meaning that people realize that in the future their benefits have to go down and/or their taxes have to go up. However, I am a firm non-believer in Ricardian equivalence, as you might have noticed when you read Lenders and Spenders.

So if I thought that economic activity consisted of spending, then I would expect deficits to increase spending and economic activity. Instead, I think of economic activity as patterns of sustainable specialization and trade, and I do not think that deficit spending is helping, because it is so unsustainable.

To put this another way, I think that long-term government bonds are fake wealth these days. There has to be some kind of default on future government commitments. There is an off chance that the future commitments that get cut will be entitlements. There is an even more remote chance that the government will find tax revenue to cover all of its commitments. We can inflate away some of our past debt, but since our projected future deficits are even higher, inflation does not make the problem go away. So I think that it is likely that we will get some sort of default. The fake wealth will be marked down at some point in the future, either through inflation or default.

Once upon a time, we had an Internet bubble that gave us dishonest stock prices. Let us stipulate that it was caused by private sector shenanigans. When that collapsed, it was followed by a housing bubble. There were private sector shenanigans involved in that one, too, but I think that some of the fingerprints on the housing bubble belong to government officials. The fake wealth that is being created now? Pretty much entirely government-generated.

Happy New Year.

EMH and Macroeconomics

Posted on by

A reader asks,

if an economist comes up with a novel and correct theory that makes predictions about macroeconomic variables, shouldn’t this theory enable him to beat the markets using these predictions?…

Therefore, it seems that if we accept both the EMH and the basic validity of macroeconomics, the latter must be about predictions that are somehow novel, correct, and non-trivial, but at the same time provide no new information about future market prices, even in terms of crude probabilities. But what would be some examples of these predictions, and what principle ensures their separation from market-relevant information?

Consider financial variables, such as the long-term interest rate or the price-earnings ratio of the overall stock market. According to the efficient markets hypothesis, these are not predictable on the basis of known information. To put this another way, you cannot beat the market forecast for these variables.

On the other hand, in conventional macroeconomics these variables can be predicted using models and controlled using policy levers. Reconciling this with the EMH has challenged economists for decades. Here are various alternative ways of doing so:

1. Policy has no effect. Markets do what they will do, regardless. The market uses the best prediction model, so economists’ macro models can, at best, replicate the market’s implicit model.

2. Policy has an effect, but markets try to anticipate policy. The expected component of policy has no effect. Only policy surprises have an effect.

It seems to me that the market monetarists (e.g., Scott Sumner) believe something closer to (2) than to (1). But (2) can get you into some strange conundrums. Does the Fed have free will? That is, does it have the ability to surprise markets, other than by acting randomly? If its actions are not random, they should be anticipated by markets. If they are anticipated by markets, then they should have no effect. etc.

I prefer a third way of looking at things, which might be expressed in the work by Frydman and Goldberg. That is, there is no reason for all participants in markets to be using the same model. They have different information sets. The EMH is a useful guide to everyone, because it serves as a reminder that it is unwise to assume that your information set is somehow superior. However, it is not correct to impose “rational expectations,” in which everyone uses the same model.

A challenge with this “multiple information sets” view of the world is that we all trade in the same market. I think that people who own long-term Treasuries and people who own gold have different expectations for inflation. Why does someone not combine short positions in bonds and gold in order to arbitrage against these different expectations? I am afraid that one has to make risk aversion and leverage constraints do a lot of work.

From this perspective, the response to policy changes is hard to predict. For both economic modelers and policy makers, a difficulty is that you do not know which models the market participants are using. Thus, you cannot know how markets will react to policy moves.

I think that such humility is appropriate. Arguments of the form “on x date the Fed did y and subsequently z happened, just as my theory would have predicted” are not persuasive to me. One can usually find other instances of the Fed doing something like y with different results.

Incidentally, I recently re-read Perry Mehrling’s biography of Fischer Black. Black was perhaps the first economist to think about the contrast between modern finance theory and conventional macro, and Black was the first and perhaps the only one to attempt to recast macro entirely in terms of modern finance. Continue reading →

The Keynesian Fixation

Posted on by

Zachary Goldfarb presents a series of charts and concludes,

Put it all together and household debt has weighed on the economy in a way many people don’t appreciate, but there seems to be progress toward addressing it.

Pointer from Mark Thoma.

In my view, Goldfarb is presenting symptoms and suggesting that they are causes. This is a characteristic of what I call the Keynesian fixation. Note that in macroeconomics the position that I criticize is the mainstream position. New readers should be aware that mine are outlier views.

To a Keynesian, economic activity consists of spending. Spending causes jobs, and jobs cause spending. When you look at economic data, that is your focus. Is spending going up or down? If it goes up, then we will have more jobs and more spending. If it goes down, then we will have fewer jobs and less spending.

The alternative approach that I propose is to look at economic activity as the formation of patterns of sustainable specialization and trade (PSST). As entrepreneurs create businesses that exploit comparative advantage and specialization, jobs and spending increase. When existing patterns of trade become unsustainable, jobs and spending decrease. See this paper and this paper.

For example, when hurricane Sandy hit New York, the Keynesian fixation was to look at the destruction in terms of how it would affect spending. Would people spend less, because of lost wealth? Or would they spend more, because of the need to rebuild? Mayor Bloomberg wanted to run the New York marathon less than a week after the hurricane hit, because he did not want the city to miss out on the all the spending that comes with the marathon.

My view is that what matters are the patterns of specialization and trade, which depend on the transportation infrastructure. I over-estimated the damage to that infrastructure, and so I thought that there could be significant long-term shifting of jobs out of Manhattan. However, if only a tiny portion of the transportation infrastructure suffered damage that could not be repaired in a couple of weeks, then the PSST story would say that employment in the area should not be affected.

To take another example, the “fiscal cliff” does not trouble me. If you hold the Keynesian fixation with spending, then it appears that a sharp reduction in government spending and/or increase in taxes will cause a recession. If you think in terms of PSST, it probably will make little or no difference. And, as I wrote two years ago,

Government spending plummeted by nearly two-thirds between 1945 and 1947, from $93 billion to $36.3 billion in nominal terms. If we used the “multiplier” of 1.5 for government spending that is favored by Obama administration economists, that $63.7 billion plunge should have caused GDP to fall by $95 billion, a 40 percent economic decline. In reality, GDP increased almost 10 percent during that period, from $223 billion in 1945 to $244.1 billion in 1947.

After the second World War, the U.S. economy easily created new patterns of specialization and trade. I think that one reason is that the war increased mobility, as soldiers met others from different parts of the country. Instead of remaining in their communities of birth, men moved in order to take advantage of new opportunities. Regardless of whether this was an important factor, it should be obvious that if we could cut government spending by 25 percent and thrive, then going over the fiscal “cliff” would be a non-event.

If I were assembling charts on the economic outlook, I would focus on secular factors: declining labor force participation among less-educated men, which indicates that new patterns of trade are reducing the value of their work; the increasing “tax” represented by the cost of providing health insurance to workers in large enterprises; the significant shift in the distribution of wealth toward Washington, DC, which makes secondary businesses such as restaurants and real estate services more profitable in the nation’s capital and less profitable elsewhere. Or consider the charts on household income and benefits from Gary D. Alexander, showing many categories of workers facing marginal tax rates of over 100 percent.

We should be trying to understand what is causing the breakdown in patterns of specialization and trade that prevailed at the end of the 1990s. We should be trying to understand what is impeding the formation of new patterns of specialization and trade today.

Housing Usually has a Positive Nominal Return

Posted on by

Thomas Alexander Stephens and Jean-Robert Tyran write,

When viewed in nominal rather than real terms, the capital gains from US housing look more appealing. From 1946 to 2012, nominal house prices showed a 12-fold increase. On an annual basis, housing investments have mostly resulted in gaining money (in 58 out of 66 years), while at the same time producing real losses more often than not (in 36 v 30 years)

Thanks to Mark Thoma for the pointer.