On the greatness of the great moderation

Lola Gadea, Ana Gomez-Loscos, Gabriel Pérez-Quirós 26 October 2015

a

A

The period of unusually stable macroeconomic activity experienced in the US that began in the mid-eighties is known as the Great Moderation. Kim and Nelson (1999) and McConnell and Perez-Quiros (2000) were the first to document the substantial decline in US output volatility in 1984, but the academic and non-academic literature on the Great Moderation is prolific, and its possible causes – namely, changes in the structure of production, improved policy and good luck – continue to be a matter of lively debate.

Although the Great Moderation is one of the most widely documented stylised facts of modern macroeconomy, two issues could question its greatness:

  • First, the Great Moderation has never before been studied in a long historical perspective as previous studies only use post-World War II data;

Can the Great Moderation be identified when considering a longer dataset? In a recent study (Gadea et al. 2015), we investigate this question. We conclude that, considering a dataset beginning in 1875, we chronologically find three structural breaks in variance that roughly correspond to the end of each of the world wars and the beginning of the Great Moderation. Indeed, the Great Moderation had as great an impact on the characteristics of the series as World War II had.  

Second, one natural question that has arisen in recent years is the possible end of the Great Moderation due to the Great Recession, which was of unprecedented severity and duration in the postwar US business cycle and led many economists to suspect a major breakdown in the GDP series. But, might output volatility not remain subdued despite the tumult created by the Great Recession? In a recent paper (Gadea et al. 2014), we formally address this question.

Great Moderation vs. Great Recession

The quarterly US GDP growth rate is plotted in Figure 1.Three challenges come to mind when we wish to analyse whether or not the Great Moderation has come to an end due to the Great Recession. First, unlike the original papers that detected the Great Moderation, we are looking for more than one break in the data. We should find a break for the Great Moderation and, if the Great Moderation has ended, an additional break associated with the Great Moderation. Second, if an additional break exists, it would show up at the end of the sample, when a structural break test is less powerful. Third, if this additional break exists at the end of the sample, the sample could be too short, implying imprecise estimates of the dynamic behaviour after the break, making it difficult to find statistically significant differences with respect to the previous periods.

Figure 1. US GDP growth rate (1953.1 – 2013.4)

We carefully analyse all these technical issues using post-World War II data. We use tests that can identify more than one break, but we find no evidence of additional breaks associated with the Great Recession, only the one of the Great Moderation. To deal with the end-of-sample issue, we propose an experiment that allows the Great Recession to start in any period from 1984.2 to 2007.3, instead of 2007.4. Regardless of where we place the Great Recession sample, the tests fail to find an end of the Great Moderation. The small sample issue is solved by generating an artificial additional sample with simulated data. We add 5, 10 or 15 years of data that have the same properties as the Great Recession data and we generate 10,000 simulations. In no case do we find an additional structural break in the data associated with the Great Recession, even when the data are extended by 15 more years (Figure 2).1

Figure 2. US GDP growth rate (1953.1 – 2013.4) and 15 years of simulated data (Great Recession sample with pre-Great Moderation volatility)

Obviously, volatility has increased during the Great Recession, but our results show that is has not increased enough to indicate a return to the pre-Great Moderation volatility. To test this interpretation, we rescale the additional sample to equate its variance to that of the pre-Great Moderation period. Even so, in most cases, the results show no additional break. At this stage, the question is why do we fail to find an additional break, even if we add 15 years of data with the pre-Great Moderation volatility?

To address this question, we generate 15 years of artificial data by bootstrapping from the pre-Great Moderation sample. Now, indeed, the results change. The most common casuistry is either that no break is found or that an additional break is found, associated with the Great Moderation. Thus, the difference between the pre-Great Moderation sample and the Great Recession period does not seem to be related only to the level of volatility.

The shape of the recoveries

To study the nature of the GM in greater depth, it would be useful to investigate whether it is equally related to each of the business cycle phases. Concerning recessions, we carry out an exercise in which we compare the growth rate of the GDP series during the last recession and the growth rate of the pre-Great Moderation data. We cannot reject that the observations of the last recession come from the same distribution as those of the pre-Great Moderation recessions.2 Regarding expansions, we investigate which features of the Great Moderation expansions make them fundamentally different from the pre-Great Moderation ones. We focus on the shape of the recoveries, as it has crucial implications for the stochastic properties of the GDP growth series, long-term economic activity and job creation capacity. Some authors claim that the peak-reverting phase of the business cycle, during which output reverts to its long-run trend, with intense job creation (as opposed to the apathetic recoveries, with sluggish job creation) disappeared after the mid-eighties. However, the severity of an episode such as the Great Recssion, unprecedented in the Great Moderation times, leaves the door open to a possible transformation of the shape of recoveries. Even though the Great Recession was powerful, and comparable to the pre-Great Moderation recessions, the first year of the expansion (the recovery phase) is clearly different from the pre-Great Moderation ones. The average growth rate during the first year exceeds the average growth rate during the second and subsequent years for the pre-Great Moderation period, but is clearly below for the post Great Moderation-period. From a statistical point of view, we show that the first year of the last expansion is different from the first year of the pre-Great Moderation expansions. However, this is not the case for the second and third year of expansions.

Conclusion

The Great Moderation is not over in spite of the Great Recession even if we use a historical dataset beginning in the 19th century. The Great Moderation was originally associated with a decrease in output volatility and was considered a great achievement in terms of reducing risk and of decreasing the frequency and depth of recessions. However, after carefully analysing the characteristics of the Great Moderation, they seem to be more clearly associated with the shape of expansions. Perhaps the benefits associated with an apparent increase in stability are paid for at a very high price. Feeble expansions may be the price to pay for low volatility.

References

Coibion, O and Y Gorodnichenko (2010), “Does the Great Recession really mean the end of the Great Moderation?”, VoxEU, available at http://www.voxeu.org/article/does-great-recession-really-mean-end-great-....

Gadea-Rivas, M D, A Gomez-Loscos, and G Perez-Quiros (2014), “The Two Greatest. Great Recession vs. Great Moderation”, CEPR Discussion Papers 10092.

Gadea-Rivas, M D, A Gomez-Loscos, and G Perez-Quiros (2015), “The Great Moderation in historical perspective. Is it that great?,” CEPR Discussion Papers 10825.

Kim, C-J, and C R Nelson (1999), “Has The U.S. Economy Become More Stable? A Bayesian Approach Based On A Markov-Switching Model Of The Business Cycle”, The Review of Economics and Statistics 81(4): 608–616.

McConnell, M M, and G Perez-Quiros (2000), “Output Fluctuations in the US: What Has Changed since the Early 1980’s?”, American Economic Review 90(5): 1464–1476.

Footnotes

1 In an interesting Vox column, Coibion and Gorodnichenko (2010) also support that the Great Moderation is not over, comparing only volatility paths. They develop the idea in greater depth in their paper where they show that the Great Moderation is not over as long as good policy has played an important role in accounting for the Great Moderation.

2 Broadening our perspective with a longer dataset, we show that recession features before and after the Great Moderation are basically equal. We find that, while the volatility reduction associated with World War II is related to changes in both cyclical phases, the nature of the Great Moderation structural break is entirely explained by the features of expansion phases, in particular, by the shape of them.

a

A

Topics:  Macroeconomic policy

Tags:  great moderation, Great Recession

Professor of Applied Economics, University of Zaragoza

Senior Economist in the Macroeconomic Analysis and Forecasting Division, Banco de España

Gabriel Pérez-Quirós

Unit Head of Macroeconomic Research, Research Department, Bank of Spain; and Research Affiliate, CEPR

Vox Talks