Curve-Fitting

I want to draw your attention to this post by Eli Dourado. In the post Mr. Dourado makes the point, too frequently neglected, that the operative timeframe for short term solutions is the short term and we’re not in the short term any more:

The case for stimulus is based in monetary non-neutrality. If we double the money supply, the real productive capacity of the economy does not increase—real productive capacity has nothing to do with monetary factors. However, because people are tricked, and because some wages, prices, and contracts don’t adjust instantaneously, output may go up briefly. Business owners see an increase in nominal demand for their products and mistakenly assume that it is an increase in real demand. They see this as a profit opportunity, so they expand production. As prices, wages, and contracts adjust to the new money supply and their assumption is revealed to be false, they cut back on production to where they were before.

If we view the recession as a purely nominal shock, then monetary stimulus only does any good during the period in which the economy is adjusting to the shock. At some point during a recession, people’s expectations about nominal flows get updated, and prices, wages, and contracts adjust. After this point, monetary stimulus doesn’t help.

He cites three different areas as evidence that the short term window has passed: corporate profits, duration of unemployment, and household debt service payments as a percentage of disposable income. I’ve seen links to this post in a half dozen different places in the econblogosphere lately and I think it’s worth a glance.

Mr. Dourado reproduces graphs from the St. Louis Federal Reserve in support of his observations. I’ll use those as the point of departure for my comments, more riffing on his post than a critique. Rather than considering the timeframe 1997 to present, the timeframe Mr. Dourado elects for his graphs, I’m going to open them up to the full duration supported in each case.

The graph illustrates after-tax corporate profits adjusted for inventory from 1948 to the present. I actually see two different curves. There’s one curve, increasing exponentially presumably with inflation, from 1948 to about 1990. After 1990 it’s off to the races with a brief decline during the recent recession.

I can only attribute that to automation indirectly. What I really think it reflects is the impact that just-in-time inventory systems have had on the return to capital. JIT systems preceded the Internet period (which I’d suggest is about 1996 to the present) and enabled companies to stock considerably less inventory. They’ve really taken hold during the Internet period. It’s truly amazing how “transshipping” has grown. Transshipping is when the end customer places the order with a retailer, the retailer places an order with a wholesaler, the wholesaler places the order with the manufacturer, and the product is produced and shipped directly to the customer by the manufacturer (or some variant of that scenario). Little or no stocking of inventories. Amazon.com is a combination transshipper and retailers service company.

I suspect that some other factors, like China’s pegging the yuan to the dollar and changing policies with respect to the taxing of capital gains, may also be operative. But I don’t see a continuous transition in operation with the present just a continuation of past trends. Something really changed in the 90s and only accelerated in the Aughts.

This graph illustrates the mean duration of unemployment from 1948 to the present. In this graph I do see the present continuing the past trend: peak mean duration of unemployment, subsequent to economic downturns, has been increasing as has been the trough mean duration of unemployment, which has occurred just before the start of a downturn. Yes, the data is a bit noisy. But I think there’s a persistent trend going back sixty years.

The greatest anomaly represented is the present. Even assuming that peak mean duration of unemployment has increased exponentially over time today’s duration of unemployment seems very long to me. Something else must be happening.

This graph illustrates household debt service payments as a percent of disposable personal income. I think I see a long term trend here, too: people take on more debt during booms and pay it down during busts. However, here, too, the period of the Aughts is an anomaly. There was no paying down of debt following the downturn of the early Aughts, only an acceleration.

I conclude with two observations. First, I think that focusing too closely on the period 1997 to the present is an error. It assumes that the early part of that period was normal while the later part of the period has not been. I would need a lot more evidence to support that supposition. Quite to the contrary I think that 1997 to 2006 is a great departure from the past while the last five years more closely resembles a return to trend.

Second, Great Moderation my Aunt Fanny. Grand Illusion, more likely.

8 thoughts on “Curve-Fitting”

  1. I’m generally not prone to slobbering adulation. However, this another great, great post.

    Let’s go to the first concept. I have for years observed and made the point that more precise inventory management and flexible employment have attenuated the business cycle. This really has been a phenomenon from the 80s. It has been credit/asset bubbles that have caused volatility. For corporations, profitability has skyrocketed.

    As for unemployment duration. Anyone here knows my views. Until we rid ourselves of the current political stance, and players, employers will stand pat. Look at the graph.

    As for the debt graph, people got suckered, they thought their 401ks and houses enabled them to take on this debt and consume. I’ve pointed out, to some claims of arrogance, that I have a significant net worth. However, my real point is that I live my life, relative to that net worth, pretty modestly. Almost a Depression mentality. Yeah, some nice cars. But the house isn’t even a fifth of what we could afford. (And what would I do with it?) But many, many people maxed out on what they could buy relative to what they could afford. That’s dumb. And now we are correcting. They got suckered and burned.

    And I think that is Dave’s point. We are returning to the long term mean. It’s time for the country to sober up. But that has two edges. Those of some financial means who overextended need to sober up. But government needs to sober up as well. We can’t pass out candy willy nilly. The party is over.

    Reality, anyone?

  2. For the first graph of corporate profits, there is only one curve with deviations due to bubbles/recessions. Take the natural log of the data and it becomes quite apparent.

    Fit that logged data to a simple trend model with and AR(1,4) process and the fit is extremely good (adj. R-square is 0.99749–i.e. this simple model “explains” 99.749% of the variation in the data).

    Or you can download the data, and then using a time variable (with 1/1/1947 = 1, 4/1/1947 = 2, etc.) and profits are described by,

    e^(2.49611+0.01837*t).

    You’ll get noticeable deviations from trend for 1992-1999 (tech bubble), followed by the recession. Then again in 2007-2007 (housing bubble), followed by the recession.

  3. That’s sort of my point, Steve. You can’t take 1996-2006 as your norm, as Mr. Dourado implies by his choice of timeframes for his graphs. The only way you can fit the bubbles into the corporate profits graph is to chop it into two graphs (before and after).

  4. “You’ll get noticeable deviations from trend for 1992-1999 (tech bubble), followed by the recession. Then again in 2007-2007 (housing bubble), followed by the recession.”

    This illustrates that for over twenty years we’ve used monetary policy to encourage leveraging and drive economic growth, the inevitable result of which is that the accelerating debt eventually overwhelms the ability of firms and households to service it. Mr. Dourado omits (Or does he? One could interpret his remarks as suggesting either that government money isn’t real money, which is ridiculous, or that he understands the endogenous theory of money) that even if the Fed is able to tempt enough people into loans to double the money supply, the increase is only temporary. Eventually the supply will contract when the loans are paid back/defaulted on.

    It’s like our national intelligentsia has brain damage, continuing to do the same thing over and over and expecting it to work this time. QE3 is entirely an act of desperation. I don’t believe for a moment Bernanke expects to make any significant difference in aggregate demand.

  5. I think the way to put it is there is one curve with deviations due to bubbles (up and down, the follow on recession). 1990 is nothing special, IMO. The real point of disconnect is 1992 when you start to rise above the trend line (so technically 1993). Right now, i.e. last several observations we are pretty close to trend again.

    This graph illustrates household debt service payments as a percent of disposable personal income. I think I see a long term trend here, too: people take on more debt during booms and pay it down during busts. However, here, too, the period of the Aughts is an anomaly. There was no paying down of debt following the downturn of the early Aughts, only an acceleration.

    You can’t make these claims based on the graph…not enough information, or more accurately you can’t disentangle possible drivers. The graph is of debt payments as a percentage of disposable income, that is,

    DP/DI.

    If DP (debt payments) goes up, then the ratio goes up. However, if DI (disposable income) goes down, then the ratio goes up as well. So After 1990 was the decline in the ratio due to increasing incomes or decreasing debt payments? I don’t know, and the graph certainly doesn’t tell us. You could even have countervailing movements in the numerator and denominator and which ever one dominates would move the ratio up or down. Again using 1990 it could be the case that people were taking on more debts and thus having higher debt payments, but that their incomes were growing even faster so the overall ratio was declining.

  6. From 1995 to 2000 median income grew by roughly five percent, while during the 2000’s incomes stagnated. Yet debt servicing as a percentage of incone rose in a relatively steady fashion until 2007. I think Steve is correct that there are more variables in play than can be gleaned from the above chart.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>