Relevant and even prescient commentary on news, politics and the economy.

The Relative Price of Housing and Subsequent GDP growth in the USA

The great recession of 2008-9 followed an extraordinary house price bubble. The sluggish was characterized by a very slow recovery of residential investment. Oddly, the extensive revision of macroeconomic models which implied a very low probability of great recessions has not involved a focus on housing. Instead it has focused on financial frictions – essentially it is assumed that the 2008-9 recession was extraordinary because a major financial crisis occurred. Dean Baker dissents (as he often does) arguing that the severity of the recession could have been predicted given the massive decline in housing prices and earlier estimates of the effect of home equity on consumption. This note attempts to being to assess that claim. It also asks if it is possible to forecast GDP growth over the medium term. Finally it is part of the Rip Van Keynes series, because I will use an empirical strategy which has been out of fashion for at least four decades – basically an ad hoc OLS regression (sometimes I even include an exponential trend).

The basic result is that if the relative price of housing is high (compared to an exponential trend) then GDP growth over the following 5 years is low (compared to an exponential trend). Aiming to test out of sample forecasting, I start using 20th century data only.

-7.52 is a fairly impressive t-statistic.
Lnindex L20 is the logarithm of the ratio of the all transactions house price index to the consumer price index lagged 20 quarters. Gdp5 is the growth of the logarithm of real gdp over the past 5 years. Quarter is the calender quarter up to the 4th quarter of 1999 = 1999.75. The data were downloaded from Fred and are described in what might be generously considered a sort of data appendix. One point must be mentioned here – the all transactions house price index is available only starting in 1975, so the first useful observation is growth of GDP from 1975q1 to 1980q1.
The series are quarterly, so the dependent variable is a moving average of changes summed over 20 quarters. In the crudest attempt to deal with this, I calculate Newey West standard errors with 19 lags. These would be valid if log GDP were a random walk with drift (the constant) and trend (growth slowdown).
This regression is at least a hint that 8 years before the great recession began, there was already evidence that extremely high relative price of housing was likely to be followed by low GDP growth. Because the regression is, at best, barely presentable, I focus on out of sample forecasting. pgdp5 is the fitted value which can be considered a very crude forecast of real gdp growth over the following 5 years.
Out of sample the forecasts and outcomes are positively correlated. The correlation of pgdp5 and gdp5 over 2000q1 through 2018q1 is over 0.86. Out of sample forecasts of GDP growth over the following 5 years seem to be quite useful. This may be simply due to the estimated trends.

The following regression shows that forecasts of deviations from trend are correlated with deviations from trend.
. newey gdp5 pgdp5 quarter if quarter>1999.9,lag(19)

This is a test of out of sample forecasting performance. It is, to put it mildly, rather more successful than out of sample tests of long term macroeconomic forecasts usually are.
The data are, perhaps, more usefully summarized with a graph. Figure 1 (finally) is a scatter of the logarithm of the relative price of housing and GDP growth over the following 5 years.

This ignores even the deterministic trends. Also the whole sample is graphed. Notably while some periods show extraordinarily high relative prices of housing and extraordinarily low subsequent real GDP growth, the GDP growth does not look anomalous. The computer is not surprised by the severity and duration of the great recession given the early 21st century housing bubble.
Here are the time series. L20.lnindexm4 is lnindex lagged 20 quarters – 4.0 (the base years for the all transactions housing price index and the CPI are different).

Notice that the first observation for the index lagged 20 quarters is 1980q1 because the index is available from 1975q1 on.
Here are the series of outcomes and forecasts. The only anomaly is that the great recession was so mild. The computer forecast 5 year gdp growth as low as -10% and it never actually was less than zero. Still this is unusually successful out of sample forecasting of medium term gdp growth.

Robustness checks etc after the jump. Also this post is available as a pdf here.

Comments (1) | |

LOMPIGHEID: “Omgekeerd omgekeerd.”

Last week I was browsing through one of the books on the shelf at work, which had in it three essays by the inter-war German Marxist Karl Korsch. One of the essays, a 1932 introduction to Capital mentioned mentioned a section in Chapter 24, “The So-Called Labour Fund” as exemplary of Marx’s critique of political economy. The “labour fund” was more commonly known as the wages-fund, the doctrine famously recanted by John Stuart Mill in 1869.

After it had been repudiated in various degrees by the economists who formerly propounded it, the defunct doctrine became a straw man “fallacy” attributed to precisely the trade unionists who had been the targets of the doctrine’s disdain. Marshall dubbed the re-purposed doctrine the fallacy of the fixed work-fund. David F. Schloss christened it the Theory of the Lump of Labour.

As is my habit, I searched on “labour fund” and “lump of labor/labour” to see if anyone had previously made the connection between Marx’s critique in Capital and the ubiquitous attributions of the fallacy by economists to non-economists. What I discovered was a six-page discussion of my own historical investigation by a Belgian economist, Walter Van Trier, published in 2013 in the Belgian journal Over-Werk.

The title of the journal is somewhat of a pun as “over” means both “about” and “above” in Dutch, so it could mean both about work and overwork in English. The word lompigheid also contains a bit of a pun — as one might guess lomp is lump and lompigheit (with a ‘t’) refers to lumpiness, while lompigheid (with a ‘d’) means rudeness or clumsiness.

I am posting below a translation of the section from Van Trier’s article that deals specifically with my analysis of the lump of labor fallacy. The full article, in Dutch, can be found here. Happy May Day!

Comments (10) | |

Job Guarantee versus Work Time Regulation

There has been a bit of commotion recently about the Job Guarantee idea (AKA employer of last resort). I don’t consider myself an opponent of the strategy but I do have several reservations about its political feasibility, the marketing rhetoric of its advocates, and its economic and administrative transparency. Some of these concerns I share with an analysis presented by Robert LaJeunesse in his 2009 book, Work Time Regulation as Sustainable Full Employment Strategy. For that reason, it would be timely to post an excerpt from Bob’s discussion of”Job guarantees versus work time regulation.”

One thing that has puzzled me about the Job Guarantee rhetoric is the invocation of Hyman Minsky as patron saint of the strategy. There is no question that he advocated a job guarantee with the government acting as employer of last resort. But in the passages I’ve read, the proposal was either contingent to a broader discussion or supplemented with various other proposals some of which might be regarded as more far-reaching and controversial even than the job guarantee.

For example, a 1968 proposal argued that, “In addition, it will be necessary to restrain profits and investments; in particular, the highly destabilizing tendency for investment demand to explode will have to be brought under control.” Nineteen years later, Minsky supported a proposal for “a maximum of 32 hours of work a week at the minimum wage” but argued it needed to be supplemented by other programs such as a universal, non-means tested child allowance. Both of these proposals were historical and context specific, with the earlier one arising from a critique of LBJ’s War on Poverty and the later one in response to Reagan administration proposals for welfare reform.

The following excerpt is from pages 125-134 of  Work Time Regulation as Sustainable Full Employment Strategy. 

Comments (3) | |

Minimum Wage Effects with Non-Living Wages

I’m teaching “Economics for Non-Economists” this semester. This is an interesting experiment, and is strongly testing my belief that you can teach economics without mathematics so long as people understand graphs and tables. (It appears that people primarily learn how to read graphs and tables in mathematics-related courses. Did everyone except me know this?)

Since economics is All About Trade-offs, our textbook notes that minimum wage increases should also mean some people are not employed. Yet, as I noted to the students, in the past several decades, none of the empirical research in the United States shows this to be true. (From Card and Kreuger (1994) to Card and Kreuger (2000) to the City of Seattle, in fact, all of the evidence has run the other way, as noted by the Forbes link.)

Part of that is intuitive. If you’re running a viable business and able to generate $50 an hour, it hardly makes sense not to hire someone for $7.25, or even $9.25, to free up an hour of your time. The tradeoff is that your workers make more and your customers can afford to pay or buy more. Ask Henry Ford how that worked for him.

The generic counterargument (notably not an argument well-grounded in economic theory) was summarized accurately by Tim Worstall in one of his early attempts to hype the later-superseded initial UW study for the Seattle Minimum Wage Study Team.

[T]here is some level of the minimum wage where the unemployment effects become a greater cost than the benefits of the higher wages going to those who remain in work.

This seems intuitive in the short-term and problematic in the long term, even ignoring the sketchiness of the details and the curious assumption of an overall increase in unemployment (or at least underemployment) if you assume a rising Aggregate Demand environment. To confirm the assumptions would seem to require either a rather more open economy than exists anywhere or a rather severe privileging of capital over labor.*

On slightly more solid ground is the assumption that minimum wage should be approximately half of the median hourly wage. But then you hit issues such as median weekly real earnings not having increased much in almost forty years, while a minimum wage at the median nominal wage rate suggests that the Federal minimum wage should be somewhere between about $12.75 and $14.25 an hour. (Links are to FRED graphics and data; per hour derivations based on the 35-hour work week standard for “full-time.”)

So all of the benchmark data indicates that reasonable minimum wage increases will have virtually no effect, and none on established, well-managed businesses. The question becomes: why would that be so?

One baseline assumption of economic models is that working full-time provides at least the necessary income to cover basic expenses. Employment and Income models assume it, and it’s either fundamental to Arrow-Debreu or you have to assume that people either (a) are not rational, (b) die horrible deaths, or (c) both.

If you test that assumption, it has not obvkiously been so for at least 30 years:


The last two increases of the Carter Administration slightly lag inflation, but they are during a period of high inflation as well; the four-year plan may just have underestimated the effect of G. William Miller. (They would hardly be unique in this.)

By the next Federal increase, though—more than nine years of inflation, major deficit spending, a shift to noticeably negative net exports, and a couple of bubble-licious rounds of asset growth (1987, 1989) later—the minimum wage was long past the possibility of paying a living wage, so any relative increase in it would, by definition, increase Aggregate Demand as people came closer to being able to subsist.

The gap is greater than $1.50 an hour by the end of the 1991 increase. The 1996-1997 increase barely manages to slow the acceleration of the gap (to nearly $1.70), leaving the 10-year gap in increases to require three 70-cent increases just to get the gap back down to $1.86 by their end in 2009.

Nine years later, almost another $1.50 has been eroded, even in an inflation-controlled environment.

Card and Kreuger, in the context of increasing gap between “making minimum wage” and “making subsistence wage,” appear to have discovered not so much that minimum wage increases are not negatives to well-run businesses so much as that any negative impact of an increase, under the condition that the minimum wage does not provide for subsistence income, will be more than ameliorated by the increase in Aggregate Demand at the lower end.

My non-economist students had very little trouble understanding that.

*The general retort of “well, then, why not $100/hour” would create a severe discontinuity, making standard models ineffective in the short term and require recalibration to estimate the longer term. Claiming that such a statement is “economic reality,” then, empirically would be a statement of ignorance.

Comments (18) | |