Macro Roundup

"...There is no contradiction in a model where both aggregate demand and aggregate supply curves shift in unfavorable directions! And in the medium run, each of these shifts pushes the other curve around too...." In the medium-run, supply- and demand-side stagnation stories are very similar

It’s not either/or. Circa 2008-2009, we should have had a higher inflation or ngdp target. We could do structural reform too, whatever you might think that means, obviously opinions will differ. We could build productive infrastructure, boosting both supply and demand, and with little risk of default on the debt. And yes, we might wish to cut some other government expenditures in the process (farm subsidies anyone?), with looser money to pick up the slack. What we shouldn’t do is all that Keynesian ditch digging. Even if you agree with the argument for it (and I don’t), it so hurts the reputation of legitimate government investment that we shouldn’t be going near it. There are many, many other better things to do, including giving our government a reputation for careful project selection looking forward.

Simple textbook economics indicates that the AD-AS distinction makes the most sense in the short run. Of course hysteresis is a mechanism by which low AD can over time translate into low AS as well, but the causation runs both ways, don’t forget that.

But if we have a persistent shortfall in demand, what we need is measures to boost spending — higher inflation, maybe sustained spending on public works (and less concern about debt because interest rates will be low for a long time). Cowen, Tyler, "In the medium-run, supply- and demand-side stagnation stories are very similar, Marginal Revolutions, October 29, 2014. Available at:http://marginalrevolution.com/marginalrevolution/2014/10/in-the-medium-run-supply-and-demand-side-stagnation-stories-are-very-similar.html

If labor force growth and productivity growth are falling, the indicated response is (a) see if there are ways to increase efficiency and (b) if there aren’t, live within your reduced means. A growth slowdown from the supply side is, roughly speaking, a reason to look favorably on structural reform and austerity.

Krugman is worried about policy implications:

Let’s not forget that economies have real wage and price stickiness in addition to nominal stickiness. That is another channel through which negative supply shocks can hurt aggregate demand and throw people out of work.

There is no contradiction in a model where both aggregate demand and aggregate supply curves shift in unfavorable directions! And in the medium run, each of these shifts pushes the other curve around too.

If there is less produced, there are eventually perceived negative wealth effects. What happens to demand? It goes down, in this case with a lag because of credit. Yes, it is a big mistake to assume Say’s Law always holds but it is an even bigger mistake to think it never holds. Supply slowdowns are bad for demand, and they likely are bad for credit creation too, which hurts demand further yet.

Paul Krugman argues the contraryhere. But let’s say there was a supply slowdown starting in 1999 or so, as reflected in wage and jobs data, masked a bit by the real estate bubble of 2004-2006 and with some of the productivity figures inflated for domestic purposes due to outsourcing.

byTyler Cowenon October 29, 2014 at 7:40 am

headshot @ Reinhart and Rogoff. the Romer's basically find that if you only look at advanced economics almost all the decline in output measured by GDP/industrial production was moderate (greater impact on GDP then IP) and often short-term (and much of the GDP impact is driven by Japanese experience) they hypothesis that the hangover from 2007/08 is largely a function of the crisis being worldwide rather then isolated to a single country.
"...This paper revisits the aftermath of financial crises in advanced countries in the decades before the Great Recession. We construct a new series on financial distress in 24 OECD countries for the period 1967-2007. The series is based on narrative assessments of the health of countries’ financial systems that were made in real time; and it classifies financial distress on a relatively fine scale, rather than treating it as a 0-1 variable.We find little support for the conventional wisdom that the output declines following financial crises are uniformly large and long-lasting.Rather, the declines are highly variable, on average only moderate, and often temporary.One important driver of the variation in outcomes across crises appears to be the severity and persistence of the financial distress itself: when distress is particularly extreme or continues for an extended period, the aftermath of a crisis is worse….We create a new, semiannual measure of financial distress in a sample of 24 advanced economies from 1967 to 2007. This measure is derived from real-time narrative accounts of country conditions prepared by the Organisation for Economic Co-Operation and Development (OECD). Our measure of financial distress is designed to capture a reduction in the supply of credit resulting from a rise in the cost of credit intermediation….These panel regressions show more complicated results than the simple summary of the 4 existing literature that the real effects of crises are severe and long-lasting. For industrial production, we find that a moderate crisis (a 7 on our scale from 0 to 15 of financial distress) is associated with a contemporaneous fall in output that is highly statistically significant, but quite modest in absolute size. Moreover, the effect starts to recede after six months, and is completely gone after two years.For GDP, a moderate crisis is associated with a contemporaneous fall in output that is again only moderate, but in this case very persistent. However, this persistence is driven entirely by the experience of Japan, which had a very prolonged slowdown in GDP growth starting around the same time as its financial crisis. When Japan is excluded from the sample, the results for GDP mimic those for industrial production. Taken together, these results suggest that in the forty years before the 2008 global financial crisis, the output declines following financial crises in advanced countries were on average moderate and largely temporary. Our source and approach do not allow us to separate financial distress arising from a decline in output from financial distress due to more exogenous factors. When we examine the behavior of the financial distress variable itself, we find that it is moderately predictable based on lagged output—suggesting that omitted variable bias may be present. Thus, our estimates are, if anything, likely to be an overestimate of the impact of financial distress. Consistent with this, we find that the results, though quite robust to most differences in specification, are much weaker when distress is not allowed to affect output contemporaneously….Figure 3 shows the impulse response functions for the two output series estimated over the full sample of 24 advanced countries, together with the two-standard-error bands. Panel (a) shows the results for industrial production, and panel (b) for real GDP. Industrial production falls noticeably at the time of the impulse to the financial distress variable. The t-statistic on the contemporaneous relationship is over 4 (in absolute value). Thenegative effect increases slightly in the half-year after the impulse. However, the absolute size of the decline following a moderate crisis is modest: the peak effect is a decline of 3.9 percent (t = -3.2). To put this number in perspective, Romer and Romer (1989) find that industrial production fell roughly 12 percent following relatively exogenous shifts to contractionary monetary policy in the United States in the postwar period. Another noteworthy feature of the response of industrial production is how quickly the negative impact of a financial crisis dissipates. The effect is a decline of just 1.7 percent (t = -1.2) within a year after the impulse and zero within two years. Real GDP also appears to decline contemporaneously with the impulse to the financial 25 distress variable. The immediate impact of a moderate crisis is a decline in GDP of 3.0 percent (t = -6.1). This negative effect grows slightly over the 3½ years following the impulse, peaking at 4.2 percent (t = -2.6). Since industrial production is normally substantially more volatile than GDP, the fact that the percentage decline in the two series following a financial crisis is roughly similar suggests that the impact on GDP is stronger. The more striking difference between the responses of industrial production and GDP is that the effects of a crisis on GDP appear to be much more persistent. Though the standard errors increase substantially at longer horizons, the point estimate of the impact of a moderate crisis is strongly negative for the full five years that we consider. …This persistence, however, is driven almost entirely by the special case of Japan. Japanese real GDP growth slowed from over 4 percent per year to under 2 percent around the time that its financial troubles began in the early 1990s, and then remained low. As a result, Japan’s experience points strongly in the direction of financial distress having large negative long-run effects….The bottom line of the focal empirical results is that the relationship between output and financial crises is not as dire as the modern conventional wisdom would lead one to believe. Using our new, real-time measure of financial distress and sensible time-series estimation methods, we find that the impact of a financial crisis on output is quite small.And, with the exception of Japan, the impact does not appear to be very long-lasting.Indeed, for industrial production, one would have to say that the negative effects are remarkably transitory….Panel (a) of Figure 5 shows the impulse response function of the distress variable to itself, estimated over the full sample of 24 countries. As with the impulse response function for output, we simulate the impact of an innovation of 7 in the distress variable (a moderate crisis). The figure shows that there is important serial correlation in the financial distress variable. An impulse of 7 is followed by a value of the new measure of 5.9 a half-year later. After two periods (1 year), the rate of decay speeds up noticeably, so that by 2½ years after the impulse any effect of distress on itself is almost gone….The small response of output in advanced countries that we find using the Reinhart and Rogoff crisis series stands in stark contrast to the usual summary of Reinhart and Rogoff as showing that the recessions following financial crises are particularly severe. Indeed, Figure 14.4 of This Time is Different (2009, p. 230) shows that the average peak-to-trough decline in real GDP per capita in a sample of crises is 9.3 percent. Some of the smaller effect we find using Reinhart and Rogoff’s crisis dates in our empirical framework is surely due to the fact that we are only considering crises in advanced economies in the post-1967 period. Most of the large declines behind that 9.3 percent average come from postwar crises in emerging economies (such as Argentina, Thailand, and Indonesia, where the declines in output were well over 10 percent) and from one crisis before World War II (the U.S. Great Depression, where the peak-to-trough decline in per capita GDP was close to 30 percent). Some of the difference also reflects the fact that we are considering the full universe of crises that Reinhart and Rogoff identify in the 24 OECD countries after 1967. The 9.3 percent is the average decline in a selected set of severe episodes. Finally, some of the difference reflects our regression-based empirical approach. Reinhart and Rogoff’s measure of the peak-to-trough decline in GDP around crises often includes output falls that predate the crisis.In contrast, the regression procedure only looks at the behavior of output following the crises.16 A third feature of the impulse response functions for output using the alternative chronologies involves the persistence of the decline in GDP. Recall that for the new series using the full sample, the decline in GDP following a crisis is highly persistent. The same is true following a crisis in the Reinhart and Rogoff series. In contrast, using the Laeven and Valencia series, there is a substantial amount of bounce back: two-thirds of the output decline is reversed within five years after the start of the crisis. In the results using the new series, the persistence of the effect of a crisis on GDP is due entirely to the special case of Japan. Japan is a large outlier in our scaled crisis series, and so the prolonged slowdown in GDP growth in Japan after 1990 gets tremendous weight in the estimation. When Japan is excluded, the GDP effects disappear within two years. In contrast, the results using the Reinhart and Rogoff series are not very sensitive to the inclusion of Japan. In their non-scaled series with many crises, Japan gets only a small weight, so excluding it matters relatively little. For the Laeven and Valencia series, excluding Japan lessens the maximum GDP decline in the impulse response function by about 20 percent, and causes the effect to go away entirely after five years. This happens even though the series is not scaled because there are only eight crises in advanced economies in the period 1970 to 2006 in their chronology—so the experience of Japan is still getting substantial weight in the full-sample estimates. The bottom line of the results using alternative crisis chronologies is that for no crisis measure is the estimated impact of severe financial distress very large. Using sensible empirical techniques on the same sample of advanced economies since 1970 shows quite minor effects on output. Financial crises just do not seem to be followed by large declines in output in advanced economies in the modern period, whichever chronology is used.17…The conventional wisdom is that financial crises are followed by large downturns and weak 44 recoveries. Yet at least for advanced countries, the prior evidence supporting this conclusion is limited. It is based on chronologies for crises that are closely related to one another; that were constructed after-the-fact by researchers who were aware of subsequent outcomes; that simplify a continuum of possible degrees of financial distress to a 0-1 variable or just a handful of categories; that combine many crises in developing countries with just a few in advanced countries; and that are often examined using very simple empirical techniques, such as comparisons of recessions with and without crises….This paper therefore revisits the aftermath of financial crises in advanced countries. We construct a new financial distress series for 24 advanced countries from 1967 to 2007 using the real-time accounts of countries’ conditions prepared twice a year by the OECD. Our new series classifies financial distress into numerous gradations, which we scale from 0 to 15. The new series displays both important similarities and important differences with previous crisis chronologies. There is agreement about the presence and approximate timing of many of the most severe crises. But we find that real-time awareness of financial distress sometimes did not come until well after the dates identified by earlier chronologies. And a few episodes that previous series identify as important crises do not show up at all in our new measure. Analysis of two other real-time sources largely confirms the accuracy of the new series. Studying the aftermath of crises using our new series and conventional regression techniques leads to a view that is very different from the conventional wisdom. Crises in advanced countries are associated with falls in output, but the falls are only moderate. When measured using industrial production, output then quickly rebounds and returns to its pre-crisis path. When measured using GDP, output does not bounce back, but this pattern is driven entirely by the experience of Japan. Interestingly, these findings are due only partly to our new measure of financial distress.…Using conventional regression techniques with standard chronologies of financial crises in advanced countries also does not provide strong support for the view that the aftermaths of crises are persistently grim. We also find considerable variation in the aftermath of significant crises. In some cases, there was hardly any departure of output from its pre-crisis path. In others, output fell below the pre-crisis path but was soon noticeably above it. And in the case of Japan in the late 1990s, output fell further and further below the pre-crisis path. One factor that appears to be important to the variation is the severity and persistence of the crisis itself. Cases where the financial distress was more limited and resolved quickly were associated with more favorable outcomes....Why was the downturn in 2008-2009 so extreme and the subsequent recovery so anemic in many advanced countries?Our finding that the fall in output following financial crises in such countries is typically modest suggests that the easy explanation for the poor economic performance—this is what always happens after financial crises—is unlikely to be true. Rather, factors specific to the 2008 episode are likely to have been key.The factor most directly linked to our results involves the severity and persistence of the recent financial distress. The 2008 crisis was the worst in nearly a century in the United States and Europe. And, although it has been over half a decade since the acute phase of the crisis, banks in many European countries are still viewed as troubled and as constraining the availability of credit. Since we find that greater severity and persistence of financial distress is associated with worse performance following a crisis, the extreme nature of the recent distress is one likely reason for the continuing economic weakness.A related factor is that the recent financial crisis was essentially worldwide, while most other postwar episodes of severe financial distress were limited to individual countries or narrow regions. When a crisis is confined to a single country or to just a few, healthy financial institutions abroad may offset some of the reduction in credit supply from domestic lenders. A worldwide crisis means that there are few healthy lenders to step into the void. More generally, a crisis in one country may have adverse spillovers to others. One would therefore expect a given rise in domestic financial distress to be associated with worse outcomes when distress also increases in other countries...."Romer, Christina and David Romer, “New Evidence on the Impact of Financial Crises in Advanced Countries,” University of California, October 14, 2014. Available at:http://www.parisschoolofeconomics.eu/en/news/d-romer-and-c-romer-new-evidence-on-the-impact-of-financial-crises-in-advanced-countries/