Recession! A History of Economic Contraction

>> John Taylor: It's a wonderful pleasure to have Tyler back to tell us about the history of economic contraction.

>> Tyler Goodspeed: Well thanks and it's great to be back with friends, former colleagues and definitely this is not an opportunity to turn down to have research on recessions. And to have the current and former chairs of the Recession dating committee.

 

As some of you may know, I'm both an historian and a card carrying PhD economist so I sometimes wear both hats or one or the other. So before getting started I actually wanted to begin with a story and please do indulge me because this is a fascinating story to me.

 

So many of you may be familiar with the recession that began in 1873. It was known for a long time as the Great Depression until an even greater one came along. And the precipitating event was the failure of the widely respected banking house of Jay Cook. Jay Cook was a really interesting character.

 

So he was born in a frontier town in Ohio. By the age of nine he was already selling trinkets out of his uncle's dry goods store. By the age of 37 he had retired worth approximately the contemporary equivalent of $100 million after a very successful investment banking career in Philadelphia.

 

But he didn't stay long in retirement because around the time of the outbreak of the Civil War he went back into the investment banking business. Not because he was looking to make more money but rather because his son and nephew were of the age at which he thought they should get some experience.

 

But it was quite fortuitous because the state of Pennsylvania was trying to move a 3 million dollar bond issue that all of Philly's bankers said you can't sell that at par. And Jay said no, I think you can sell that at par on a wave of patriotic zeal.

 

Sure enough he did. And he sent a letter to Jefferson Davis, the President of the Confederacy saying that the north millions stand behind the effort against this rebellion. He ends up selling about a quarter of the federal government's debt during the war ends the war not just wealthy, but fantastically so.

 

He buys a palatial mansion north of Philadelphia and a castle on an island in the Great Lakes where he's starting to spend more time. When he was briefly working in a dry goods store in St. Louis as a teenager he'd heard firsthand the stories of Clark. And there's a new railroad project, the Northern Pacific railroad that aspires to connect the Great Lakes Duluth to wonderful natural harbor in Puget Sound in the Pacific, Seattle.

 

And crosses some of the most fertile agricultural land in the country, if not the world abundant natural resources. And Jay is impressed by the recent completion of a survey of the Yellowstone River Valley. And he starts lobbying the grant administration to make a national park cuz he sees tourism potential.

 

Now the problem for Northern Pacific is that this was gonna be entirely privately funded. The more famous Union Pacific and Central Pacific, they had benefited from federally subsidized loans. Northern Pacific was gonna be entirely private. And after commissioning some surveyors and lawyers January 1, 1870 Cook inks the deal.

 

He's gonna be the sole underwriter for this massive undertaking. And it initially goes pretty well. So they break ground for the first time. They finish a connector line from Duluth to St Paul. But then challenges start to mount so.

>> Michael Boskin: A contemporary standards of California's high-speed rail.

>> Tyler Goodspeed: Yeah, he was doing pretty well.

 

So the surveying teams start to trespass on traditional indigenous hunting grounds and occasionally come under attack by forces of the legendary Sitting Bull. His forces never numbered that very never numbered many. But the press covered these attacks in lurid detail in part because the flamboyant George Armstrong Custer was providing coffee.

 

He then had, in the fall of 1872, just as the Grant campaign for re election was in its closing stage. The New York sun breaks this incredible scandal of corruption at the Union Pacific where basically the directors were attempting to defraud investors in the Union Pacific Railroad casts a huge pall on US Railroad securities.

 

But then there's an additional complication and it's the arrival in St Paul of this man. He had an aristocratic bearing and locals noticed that he had his harness. Horse's harness bore an earl's crest, and his signet had the initials. His monogram had the initials GHG, which were the initials of the recently lost Earl of Aberdeen Lord Gordon of Aberdeen.

 

He had gone missing at sea and he deposits 20 to $40,000 in a local bank. And lets people know that he's in the market to buy about 50 or 60,000 acres of land along the future route of the North Pacific to relocate the tenants from his overcrowd Scottish estates Northern Pacific.

 

They immediately leap to action. They organize a surveying party for him. White glove service, French champagne, a mult long wagon train. Spend the equivalent of millions on this. And they even fly the banner of the clan Gordon. He leaves at the end of 1870, has selected tracks but never executes the transactions.

 

Goes to New York, gets some fancy rooms at the Metropolitan Hotel then owned by Tammany Hall Boss Tweed. And he conveys to the embattled president of the Erie Railroad, one Jay Gould. Some of you may know him for having brilliant financier. He orchestrated the failed attempt to corner the market for gold few years before.

 

And he's trying to battle off some English investors who are saying we want higher dividends and you need to stand for re election each year. And Lord Gordon Gordon says, I control directly or indirectly 60,000 shares in your railroad. And I'm willing to back you if you make a few reforms.

 

But in the course of my investigations, I've incurred about a million dollars worth of costs. And I think it's reasonable to expect that I'd be reimbursed for that at least in half. Gould cobbles together $500,000 in railroad securities and cash hand it to Gordon Gordon asks for a receipt.

 

Gordon hands it back and says, my word should be enough. My honor should be enough. Gould accepts that. A couple weeks later, he learns by telegraph that Gordon is trying to sell the securities in Philadelphia. Instead of going to the police, Gould then brings the police commissioner of New York City and a judge and Tammany Hall Boss Tweed to the Metropolitan Hotel and says, if you don't return these securities and the cash, I'm going to haul you to Ludlow street jail.

 

Gordon Gordon returns it, then tries to countermand the cancellation of the security sales. Gould gets a warrant for his arrest. There's gonna be a trial in the middle of the night. Gordon Gordon absconds to Canada, where after an extrajudicial extradition, is eventually located by Canadian magistrates in Winnipeg for some jewelry theft in Edinburgh and London.

 

Before they can extradite him, in the words of the contemporary press report, blows his brains out. The reason I tell you all of that is because the press at the time covered this blow by blow, especially in Europe. And so you had editorials and letters to the editor asking what was going on in the US railroad industry.

 

And they even noted that how is Gould still being sought out at all, given his reputation? And they wrote, only when those who still control the largest railway interest in the United States are no longer found to flock together as one of a feather as Jay Gould, might it become possible for distant shareholders to regard their American investments without distrust.

 

So by summer 1873, Jay Cook, who had been hoping to sell $100 million in northern Pacific bonds in Europe alone, they're going days without selling a single bond. And in the meantime, he's extended a million and a half in overdrafts in the Northern Pacific because they need to build track to sell bonds.

 

They need to raise money by selling bonds to build track. He's got millions of dollars in unsold and increasingly unsellable bonds on his books. And for every previous tranche of bond sales, he acquired equity in Northern Pacific. So he's got his entire asset side of his balance sheet is frozen.

 

And then this happens. In summer 1873, a strange cloud appeared in the American Midwest and plains. And actually the pioneer children's author Laura Ingalls Wilder wrote about this. She noted a strange cloud unlike any she had seen before approaching. And it sort of shimmered in the sun. And then something hit her head.

 

 

>> John Taylor: Climate change.

>> Tyler Goodspeed: And it was a grasshopper. And Laura Ingalls Wilder, Laura Ingalls then, realized that the cloud was hailing grasshoppers. In fact, the cloud was grasshoppers. Now, this was a devastating shock for a railroad whose main pitch the bond buyers was that this is great settler community.

 

And settlers who had already settled in the the greater Midwest and Plains area, started writing back to Norway, to Ireland, to Germany. They sent letters noting what was happening. And by the way, this shock came back in 1874, came back in 1875, came back in 1876. One swarm in 1875, Albert Swarm was larger in area than the state of California.

 

And actually so key to this bond pitch of attracting European settlers was that they named the capital of the northern Dakota territory, Bismarck. So by September of 1873, the cook's partner and head of his New York office gathers some of the biggest bankers in New York together and he says, I need a million dollars in cash now.

 

They said, what can you offer for collateral that's not Northern Pacific? He says, nothing. They say, no deal. And within minutes, he closes the doors. Philadelphia and Washington have to do the same. Dozens of banks immediately suspend payment. And by the way, at the same time, United States has just demonetized silver.

 

Germany demonetizes silver, so you get a deflationary impulse. But what is fascinating, despite all this having to go wrong and go wrong in a very chance way immediately, Cornelius Vanderbilt says, well, this is the folly of building a railroad from nowhere, through no man's land to no place.

 

And even after the railroad is eventually completed, so this is the golden spike ceremony, they are lampooning this as a charlatan, huckster endeavor. So you can see a genuine English lord, they're trying to attract these sober investors. They're advertising a genuine English lord, a Bavarian general who looks a lot like Bismarck, a caricature of a German baron.

 

And the trombonis there is playing from the New York Sun. If you read Kindleberger, say, or any number of encyclopedias of US Business cycles or depressions, you would probably. By the way, before I get to that, do you know who owns the Northern Pacific Railroad today? Warren Buffett, most legendary value investor of all time.

 

I mean, so if you read Kindleberger Manias, Panics and Crashes, you would probably surmise that railroads were serial killers of economic expansions in the 19th century United States and United Kingdom. 1837, 1857, 1847, 1873, '69, '87, '93. But if you look at the cumulative mile of railroad track in the United States and United Kingdom, it is remarkably fateful.

 

And there is no instance in which a few years after recession, the cumulative mileage doesn't catch up to where it would have been had it continued uninterrupted along, along trend. So those slowdowns, no stops. This is not a stepwise function. The slowdowns during recessions are negative deviations from trend.

 

The preceding expansions are not positive deviations from trend. I think this speaks to one of the central theses of the book, which is that we are pattern seeking mammals. It's how we process incoming stimuli and match those features to patterns that we've stored and encoded in our memory.

 

So it's called feature analysis. And it's part of why many of the early psycho theorists, William Petty, Clement Jugular, were trained as medical doctors, and they would discuss economic fluctuations in medical terms, crisis, malady, remedy. And Kindleberger, he wasn't trained as a doctor, of course, but he would refer to biologic regularity.

 

And this process of pattern recognition has served us very well as a species. I mean, it's why we don't eat colorful mushrooms, it's why we don't drink stagnant water, because it often helps us to correctly identify genuine patterns to observed phenomena, and we call those epiphanies. But this pattern seeking nature can also lead us to identify, falsely identify incorrect patterns, or else to assign patterns.

 

To phenomena that are fundamentally random. So when Kindleberger was born, the year he was born, there was great fear of fire and plague. Because the last time Halley's comet visited, we got fire and plague. Now one apophany that, and we call those apophanies, in contrast to epiphanies. Now one apophany that-.

 

 

>> Speaker 4: Is that a word?

>> Tyler Goodspeed: It is, yes.

>> Speaker 4: Okay.

>> Michael Boskin: Can you spell apophany?

>> Tyler Goodspeed: I can, although I couldn't before I started this. But one apophany that we might have thought Diebold and Rudebusch drove a stake in was the notion that as an expansion ages, excesses or imbalances or misallocations accumulate, rendering the probability of recession more likely.

 

But I can confirm in the post depression Diebold and Rudebusch were looking at post war, I view post war as a bit of an arbitrary demarcation. But post depression, I can confirm their finding that the marginal effect of an additional quarter of expansion age on the probability of death is statistically no different for a 10 year old US expansion versus a 1 year old US expansion.

 

And that is even more pronounced in the United Kingdom where the point estimate of that slope is basically zero. But Diebold and Rudebush were only half right because they said that this was a phenomenon of the post war US economy but that pre war US economies did die of old age imbalances accumulated.

 

Well that wasn't true for the United Kingdom. Both pre-depression and post-depression UK expansions were statistically not likely to die of old age. And actually the reason these lines truncate where they do, so at 42 quarters for the United States and 104 quarters for the United Kingdom, is those correspond to the longest expansions observed.

 

And we're gonna come back to this. Over the past 325 years, the United Kingdom economy has been vastly less recession prone than the US economy.

>> John Cochrane: Then why isn't it 100% at the right hand?

>> Tyler Goodspeed: So it converges, cumulatively it converges. But I just truncate it there because we haven't, it goes on.

 

You can extend it out, but I just truncated there for visual purpose.

>> John Cochrane: The only longest expansion we've seen is 41/4. Then conditioned on expansion lasting X, you have one data point and it must be that you had death at 41 quarters. You can't do 42 quarters. Misunderstanding how you make a graph.

 

 

>> Tyler Goodspeed: So I just estimate a clog log here that's regressing the probability of death on age.

>> John Cochrane: Okay, yeah, parametric is the answer, okay.

>> Tyler Goodspeed: So UK expansions have never died of old age. Now the Diebold and Rudebusch's finding that pre-depression US expansions did die of old age, I can confirm that.

 

I mean you see that line is very sharply upward sloping and I can reject the null that it's zero. But that's using the official MBER series, which is Christy Romer has demonstrated before 1927 and Joe Davis has demonstrated before 1927. That probably overstates the amount of time that the US was in recession.

 

If we substitute in the Roamer dates then that slope is zero. The intercept is higher because US expansions were shorter lived pre-depression, but the slope is zero.

>> Hoyt Bleakley: Imbalances accumulated in the pig iron production or whatever was used to impute the recessions.

>> Tyler Goodspeed: I mean, no imbalances accumulating.

 

No, cuz if you use her dates, pre-depression US expansions didn't die of old age either.

>> Hoyt Bleakley: The old set of imputations run in parallel set.

>> Tyler Goodspeed: Right, right, right, yes, I see what you mean.

>> Hoyt Bleakley: Yeah, that's a way of weighting based on data.

>> Tyler Goodspeed: Yeah, correct.

>> Michael Boskin: And the data are just complete.

 

I mean we only had national income accounts starting with prisons in 30s. It went back sort of, but.

>> Tyler Goodspeed: Yep, and what's interesting, you can actually go back on an annual basis to 1700. US expansions were likely to die of old age according to the using the old NBER series going back to 1790, this was the Moore and Zarnowitz extended series.

 

But if you substitute the new Davis series, it actually slopes down. Statistically it's no different from zero. The UK also slopes down, which suggests, if anything, UK expansions were less likely to die the longer they lived. It turns out that's because of a correlation with war. Because basically the longer sort of 18th century UK expansion lived, the more likely it was in a time of peace.

 

So the probability of recession was highly correlated with being in war and peace was highly correlated with being an expansion. So if you correct for that, you get a modestly upward sloping line for the UK, but it's statistically no different from zero. Basically US and UK expansions have never died of old age.

 

What about recessions? Well, like Diebold and Rudbush, I can confidently confirm that post depression US recessions have died of old age, but that is not true for the United Kingdom. So Diebold and Rudebusch said the age dependence of post-war recessions is a function of institutional changes since the 1930s and expanded state.

 

Well, the United Kingdom had that as well. But their recessions post depression seem to be pretty age independent. Diebold and Rudebusch also found that pre-war US recessions didn't die of old age. Again, they attribute this to institutional changes since the 1930s that shortened recessions. But again this is a function of the old MBER chronologies.

 

So if you use actually if you just take out the misidentification of the 1873-78 recession, which the official chronology lists five years but the revised chronology says there's only two. If you just take that out then pre depression US recessions also died of old age. You can confidently reject that null that the slope is zero.

 

And for the UK pre-depression, UK recessions died of old age and post-depression UK recessions. Actually if you strip out a very unique recession in 1943 to '47 at the tail end of World War II, just stripping that outlier, then post-depression UK recessions also died at old page.

>> John Cochrane: Basically the standard definition of negative growth rate as opposed to level being low.

 

Also a way of thinking of when is the level low? How long does it take for the unemployment rate to cross back to where it is? My impression is those things are even more Poisson processing than the growth rate definition.

>> Tyler Goodspeed: Yeah, I think that's fair. I think that is fair.

 

My key takeaway from all of these graphs is that fundamentally expansions have never died of old age. In contrast to Diebold and Root Bush and recessions have always died of old age.

>> Michael Boskin: Getting back to John's point, you can have very, very slow recoveries that take many years to get back to the pre-level like after the financial crisis, Obama expansion, it was very, very slow.

 

Slower since World War II things bounce back quickly. And Milton Friedman had the violin analogy, etc. But are you gonna tell us what you think is going on here?

>> Tyler Goodspeed: I am getting to that, yep. And we might also ask the question, do longer recessions follow longer expansions?

 

And the answer to that is an unequivocal no. There is nothing about the duration of an expansion that tells us anything about the duration of the subsequent recession. If anything, post depression US Recessions, post depression shorter recessions might follow longer US Expansions. I wouldn't place much stock, I mean it's not a big sample size.

 

So basically, no, there's nothing about expansion duration that tells us about recession duration.

>> Michael Boskin: You can get to the kinda more fundamental point about maybe that Lucas would make if you would go live at this, looking at the longer term growth. It'd be how much cumulative growth there was and comparing that to how much cumulative reduction there was coming back, et cetera, rather than just sharp turning points.

 

 

>> Tyler Goodspeed: You have jumped pretty much to my conclusion, but to the-

>> Michael Boskin: Lucky guess.

>> Tyler Goodspeed: Yeah, but to the plucking, yeah, I mean there is a question. Well, is there something about the cumulative trough to peak percent change during an expansion or the trough to peak rate of growth during an economic expansion?

 

Is there any information in that that tells us about the peak to trough decline or the pace of the peak to trough decline during the subsequent recession? If you look at rows 1 and 5 for the US and UK no, there is absolutely nothing in the trough to peak percent change that can tell us about the peak to trough percent change in the subsequent recession.

 

And check out some of these R squares, it explains none of the variation. But the peak to trough percent change during the recession can explain a lot of the variation in the trough to peak percent change in the subsequent recovery. And actually with Wald tests, I can't reject the null that slope is actually negative one United States.

 

There are some issues with the United Kingdom in terms of. There are a couple recessions in the early 18th century that violated the fucking model cuz there were wartime recessions with big, big harvest shocks. So sort of quarter to 25% decline in agricultural output in 1705, 20% decline in 1709.

 

But outside of those two, similarly UK, UK trough to peak increases are very much explained by the preceding peak to trough decline.

>> John Cochrane: It's actually very relevant to an old literature on random walks and unit routes. And also the question of is it technology shocks or aggregate demand?

 

You're giving a very conventional story. There's a potential output, we fell down below and potential is up there. We'll get back to where we were. Very much a trend stationary shock in the downward direction.

>> Tyler Goodspeed: Yes, and I'm familiar with the Campbell and Mankiw paper and a better Cochrane paper, yes.

 

 

>> Jon Hartley: And this is consistent with Milton Friedman's fucking findings, right? And he's got that old while since I've read it, Economic Inquiry article. I think the UK is included. He had a few countries in it, but it's the idea that stronger booms don't necessarily predict deeper recessions, but deeper recession predict stronger subsequent boom.

 

 

>> Tyler Goodspeed: This is consistent and it's not just deeper recessions are generally on average followed by steeper recoveries and there are also faster recessions on average are followed by faster recoveries. But again this is on average and there are exceptions, particularly when there's regime change. And I would contend that 2008, 2009 there was, there was a regime change.

 

 

>> Michael Boskin: Another thing that's going on over this series is governments are growing and the automatic stabilizers become a much larger feature of the economy.

>> Tyler Goodspeed: So we're gonna get to that, yeah.

>> John Cochrane: The point estimate is kinda half permanent, half transitory.

>> Tyler Goodspeed: Yeah, I mean as I said, a wall test can't reject negative one, but.

 

 

>> John Cochrane: Also can't reject negative 0.5.

>> Tyler Goodspeed: Right, point taken.

>> Valarie Ramey: There's a recent paper by Beraha and Wolf and they have a model that recessions where durables are particularly hit are followed by recoveries that are stronger. And then we actually use this notion in a recent paper on the end of World War II.

 

And the idea is that if your durable stocks go down, then you get this investment boom coming out of the recession.

>> Tyler Goodspeed: Yeah, no composition does matter. And so this is real GDP going back to '70, '90. I mean we see similar patterns with real investment going back to 1901 in the US and 1830 in the UK.

 

Similar pattern with bank lending going back to 1880 in the UK. 1830 in the United States, not much of a plucking pattern in real consumption. And I think that's because there's just less variation in consumption. It's so-

>> John Cochrane: Consumption-

>> Tyler Goodspeed: Inertial.

>> John Cochrane: Forward looking as the present value, future income, it tells you the permanent component, all those great things about consumption, right Bob.

 

 

>> Tyler Goodspeed: And I also ask because Reinhart and Rogoff tell us that recoveries that follow recessions that involve financial crises are slower. The estimated coefficient, this is for the United Kingdom going back to 1700, coded for whether there was a financial crisis during the recession. Those estimated coefficients are zero.

 

And I also substitute the peak to trough decline in lending, bank lending during the recession. And the estimated coefficient on that is actually negative. So if anything, bigger peak to trough declines in bank lending are followed by bigger trough to peak percent changes.

>> John Cochrane: The usual story is not so much the age of an expansion causes recession, but overbuilding, over investment, over lending too much credit.

 

Everything they said at for Coolidge.

>> Tyler Goodspeed: Yeah, so on the overbuilding that was the investment. Yeah.

>> John Cochrane: You just showed it.

>> Tyler Goodspeed: Yes, and also all these interwar theorists, they postulate that there are imbalances between investment and consumption. So here I ask is there anything about the duration of the expansion or the trough to peak percent change in the investment to consumption ratio that tells us about subsequent changes in the investment to consumption ratio during the recession?

 

No, it doesn't tell us anything. But the peak to trough percent change in that ratio and the duration of the recession do tell us something about the subsequent change in that ratio and about real financial crisis.

>> Michael Boskin: Michael Bordeaux has a paper exactly downstream.

>> Tyler Goodspeed: And then this is real non residential structures because, yes, scraper index.

 

A lot of people say this is particular to structures, to building things. But if anything, a bigger trough to peak percent change in real non residential structures investment is followed, is followed by shallower contractions and whereas a bigger and a bigger annual rate of growth is followed by a slower contraction during the recession.

 

 

>> Speaker 9: You may get into this later but how are events going to factor into these growth line of recession?

>> Tyler Goodspeed: That's a good question and actually that's why I included this because we often confuse stocks and flows. So you read even very careful scholars, I won't name names but super careful scholars and myself included will, will invariably slip into the language.

 

Well there was a boom, there was a western land boom, there was a railroad boom. And we kind of do it in a hand wavy way. And there's a lot of this about the 1920s but remember this is the green line is the flow of building activity. And there was a war, there was a post war recession, a pandemic and a wave of strikes in 1919 and 1920 including a major coal strike in 1920 in the United States.

 

So the 1920s were about the flow returning to trend because if you look at the cumulative level of building activity it was just slowly getting back to, gradually getting back to trend after that long shock because those were real events. And I think we often cognitively forget about that.

 

I'm somewhat conscious of time and I don't think you guys will need any convincing that kitchen jugular Kuznets Kondra FTF cycles. There's nothing informative about lags of past recessions on the probability of recession today. Actually I do have to note I've been mentioning a lot of interwar theorists as all of these western interwar theorists were trying to postulate endogenous or self generating theories of business cycles.

 

There was a Soviet economist at Kondratievsk Institute in Moscow who, Eugene Slutsky, we all know slutsky. He concatenated ten digit Russian lottery numbers and constructed moving sums. And that summation of, of random and in mutually independent chance causes maps very closely onto an undulating cycle of, of UK business cycles in the 19th century.

 

So it's a Soviet economist that did that. And actually I probably don't need to convince you guys about just the fundamental unpredictability of recession. So number of false positives for US yield curve spread, UK is littered with false positives and false negatives. And if you go back before the 1960s, yield curve inversions are a banality.

 

So before the period of high inflation, the 60s, 70s, 80s, yield curve inversions are a banality.

>> Michael Boskin: It was Paul Samuelson's script that stock market predicted nine to last five recessions.

>> Tyler Goodspeed: We usually think of bond markets are smarter than equity markets but they can be wrong too.

 

This is getting to your question, Mike, which is okay, so they're unpredictable. There's not much in the expansion that can explain variation in the recession, but there is variation in the recession that can explain variation expansion. But can we at least ask sort of how have the contours of expansions and contractions been changing over the long run?

 

And here I use the improved Davis annual chronology instead of the official nber chronology before 1927. And you can see with the possible exception of a very long 16 year expansion that ended in 1856, there's a gradual upward trend in the age of US expansions and similarly an upward trend in the age of UK expansions.

 

But there is no sharp break in 1913, in 1933, in 1945. If I do a sequential test for an unknown number of breakpoints at unknown points, maybe 1807 in the United States. Maybe 1769 in the United Kingdom, but no sharp breaks in the 1930s or 1940s, when we think of these institutional changes as having fundamentally altered these, the.

 

 

>> John Cochrane: US the last one, two, three, four kind of stand out. I can't really figure out what the date is that 1990ish 1980ish, after which.

>> Tyler Goodspeed: So the long-.

>> John Cochrane: 60s.

>> Tyler Goodspeed: Yeah, yeah. So the longest expansions on record were those that ended in 2019, in 2001 and 1969 and 1990.

 

And you can see here already, I mean, how much more durable UK expansions have historically been.

>> Michael Boskin: If you look back at the previous, back to that and we look at the period before the long expansions, there are a lot of recessions and they often occurred quite rapidly on top of each other.

 

 

>> Tyler Goodspeed: Yes.

>> Michael Boskin: That can ever be part of the story or.

>> Tyler Goodspeed: Yeah, so actually I'll get to the regression results. I could just stare at these ones. I mean they look like barcodes, but I could contemplate these endlessly. These are recessions. Its recession age at the time of death going back to 1700.

 

And what really stands out to me. So you have of course, the Great Depression, which is a lot of shocks and maybe we can get into those shocks in the discussion. The four distinct banking crises, three of which were regional. This is revolutionary War. Revolutionary War, this one was fascinating for me.

 

So this popped out in the statistics available. Statistics like money supply, some price data, bills of exchange rates, but above all, trade. It is a three year collapse in trade in the colonies and there was no major war going on. I was like, what is going on here?

 

Pirates. So this is the peak of the golden age of piracy. And Edward Teach actually successfully blockades, better known as Blackbeard, successfully blockades the port of Charleston. They're raiding as far north as new Newfoundland. They're intercepting ships going from the Caribbean to Salem, Massachusetts off the coast of Long Island.

 

So trade just grinds to a halt. But those are the longest recessions in US History. So piracy, slash, call it war of the time of a type, war, great depression. This is war and some very severe environmental shocks. Some really bad winter frost followed by spring flooding and then a bunch of war related things.

 

This is post war. Well, it started sort of end of the war and then you have pandemic, you have a wave of strikes and then this is that exceptional recession that I mentioned that started in 1943, continued after the war as they expanded ration. So the ration was actually stricter after the war than during the war.

 

And then 46, 47, you again get a very extreme winter event just after they've nationalized the coal industry. And so the coal is freezing at the pits and you have 75,000 railroad cars that are frozen at the depots. Coal ships are flashed to the coast. And then as in some of these 18th century shocks, come spring you have this frozen ground because the system had settled over the British Isles.

 

Well, actually it was over the North Sea for over a month. Come spring the ground is still frozen and then they get torrential rains and they get flooding. So these are very exceptional recessions.

>> Valarie Ramey: I was looking at the UK unemployment statistics to see what happened after their war and their unemployment rate didn't go up much.

 

So, I'm surprised given that they were still in recession. It was the end of the war that you don't see it in the unemployment statistics. Have you looked at that?

>> Tyler Goodspeed: I have and I think it's in part because, I mean there wasn't as much demobilization as in the United States.

 

They still have a lot of insurgencies going on. And then also you started to have the nationalizations.

>> John Cochrane: You could have collapse.

>> Tyler Goodspeed: So yeah, before the end of the war, Clement Atley takes over.

>> Speaker 4: At the end of the war.

>> Tyler Goodspeed: Yeah, Clement Atlee becomes Prime Minister Churchill '45.

 

 

>> Robert Hall: Can I ask a question beyond the immediate history? Is there a policy implication in here?

>> Tyler Goodspeed: Yes, and I would say we shouldn't look to policy to prevent or end recessions. We also shouldn't look to policy to let recessions play out. Apply palliatives because the suffering is real.

 

But do not look to government to either end recessions or prevent recessions.

>> Robert Hall: Okay, so can I summarize that welfare but avoid overstimulus.

>> Tyler Goodspeed: Yeah, stimulus is not going to end a recession.

>> Robert Hall: Okay, so we are just looking on.

>> John Cochrane: Are you in recession or not? It kind of raises the question is there such a thing as recession or is this just a time series?

 

Is above or below zero a distinct phenomenon in a meaningful sense? And of course how deep are the recessions? They used to be deeper.

>> Tyler Goodspeed: I think it's meaningful in the unemployment sense as this room knows better than any. You would do a very good job picking out the NBER chronology by just looking at a graph of the spilling unemployment rate.

 

 

>> Michael Boskin: The substitution of a-

>> John Cochrane: Statistically meaningful, not. Personally meaningful, but just does a decline in GDP down to a tenth of a percent growth. Is that a different phenomenon than the decline that crosses zero and becomes a recession? Is the unemployment spikes smaller than they used to be?

 

Did automatic stabilizers actually do any good? Those are questions that you can't just do by counting. Are you in recession or.

>> Tyler Goodspeed: Yeah, no, that is true. I do, however, again to get at that question of how is policy attenuating these episodes of contraction. So here I just regress recession time, recession age on time and also separately on a post war dummy and recessions.

 

There has been no effect of time on the average duration of recessions united the US and UK and I mean those R squared, they're just basically zero. The post war dummy using the NBER series before 1927. That does seem like they've the recessions have gotten shorter in the US but not if you substitute the Davis Romer chronology.

 

Whereas expansions have been getting have been living longer over time and yes, post war as well. But the post war effect is basically a function of just the time. There's no discontinuity but we have been getting better at absorbing the kinds of shocks that historically would have triggered recession.

 

 

>> Alvin Rabushka: And one more question. Pre gold standard, post gold standard make any difference?

>> Tyler Goodspeed: Well, it's hard to say because of this graph and actually the next four. So these-

>> Alvin Rabushka: Money fixed to gold and money free floating.

>> Tyler Goodspeed: Right, so this is within cycle variance for real GDP, the unemployment rate and then I go to consumption and investment going back all the way to 1700.

 

Well real GDP, yeah, real GDP back to 1700 for the UK cuz I'm not interested in growth in the volatility of positive growth that when we're an agricultural economy we go like this, but it's positive. I'm interested in the volatility across positive and negative growth during sequences of expansion and contraction.

 

So yes, some people might have found that since the first half of the 20th century, volatility of real GDP in the United States declined. But that was the aberration. It declines to levels that were more frequently observed in the, in the 19th century. And by the way, this is a period, as you know, of the Great Depression and also a couple world wars.

 

This was Napoleonic wars for the UK and a lot of wars here for the uk. Somewhat similar story for the unemployment rate going back to 1765 for the UK and 1890 for the US, this is the aberration.

>> Speaker 12: And following up on John's observation, what if you were simply trying to estimate the variability in the growth rate of the US and UK economies and you forgot whether it's recession or not.

 

That's a dichotomous variable. Anytime you introduce a dichotomy into variation over time, you're going to get some funny things. But if you just estimate variation, what would you get?

>> Tyler Goodspeed: The it's just-

>> Speaker 12: Yeah, forget about whether it's a recession or not. Just look at the variability.

>> Tyler Goodspeed: So just variability over time has generally declined with the transition from primary to secondary to tertiary production and also with changes in banking that we've become financially more sophisticated.

 

And I'll actually have a few graphs about just how underbanked the United States historically was. But yeah, I think overall volatility, not taking out an expansion or contraction has generally declined over time.

>> Michael Boskin: That's called the great moderation. So it wasn't.

>> Speaker 12: Affected by how much growth rate there was at one time.

 

I mean, you're sort of saying expansion doesn't predict a recession, but does expansion predict a slower growth rate?

>> Tyler Goodspeed: No.

>> Speaker 12: Well, it's saying no, but.

>> Tyler Goodspeed: Except insofar as the deeper you get into a recovery, the harder it does become to add more jobs. The harder it becomes to sort of fill unfilled contracts, basically because there's just less slack.

 

 

>> John Cochrane: Did you tell all this to Trump?

>> Tyler Goodspeed: This has been researched since 2020.

>> John Cochrane: To Alvin's previous question is very interesting on the gold standard versus current arrangements cuz part of the common mantra. I think you're gonna find that prices were much more volatile in the olden days and part of the common mantra is, prices are sticky downwards on and so forth.

 

So that should have been bad for recessions back then. If you're finding it wasn't, that kind of overturns a very common nominal stickiness is really important for recessions and getting away from the gold standard. So you didn't have to have deflations that's very important. Maybe not.

>> Tyler Goodspeed: Yeah, I think that's generally right, one channel.

 

Was that when you had adverse harvest shocks, whether it was cotton, mostly cotton, is that you would get a big decline in net exports, which would lead to external external drains, which could have a deflationary impulse. So there's kind of this irony that you would have a lot of supply shocks accompanied by deflation, cuz there were also internal drains.

 

It's because you would have rural depositors withdrawing their deposits from local unit banks that only had one branch. So you would get internal drains as well. So it's kind of this irony that you have a lot of supply shocks in history that are accompanied by declining inflation in the aggregate.

 

Yeah, Hartley.

>> Jon Hartley: Pre-1930, post-1930, obviously, fewer recessions, longer expansion happening, how much do you think this being the result of the dearth of banking crises post? One theory would say, well, we introduced deposit insurance in the 30s and we don't really have banking crises anymore. I mean, we had a run in 2008, but that wasn't a traditional bank run, that's another kind of thing, we fixed it with capital requirements.

 

But that kind of explains this sort of longer runtime. I mean, cuz I think there were a lot of recessions pre-1930s that were just the banking crisis, 1907. I forget the exact years, but how much would you attribute to maybe a dearth of banking crisis?

>> Tyler Goodspeed: That is a very important question and a very good hypothesis because it's consistent with mine.

 

And it's actually one of the key reasons that I look at both United States and the United Kingdom. One is because those are two economies I know, I don't know the economic history of France or Germany very well. I do know those two very well, they have some of the best data extending back a long way.

 

And also, I think it's really interesting to compare two similar economies that are very deeply integrated, that have similar legal institutions, and yet have very different recession experiences. And I'll get into this, one of the reasons that UK was much less recession-prone was precisely because of the differences in banking.

 

So from 1825, you have nationwide branch banking in the United Kingdom, and a big national economy is by and large a pretty diversified portfolio. And whereas, as many of you know, you couldn't bank across state lines for most of US history, and even within states, you couldn't operate more than one branch.

 

So I will get into that a bit more as well.

>> Michael Boskin: So the development of national markets proceeded later in the US ability for the overall economy to balance things out. People move from the Midwest to Texas, etc., internal mobility offsetting some of these issues, etc.

>> John Taylor: Michael Bordo.

 

 

>> Tyler Goodspeed: Great, Mike.

>> Michael Bordo: Tyler, I read this book and it's fantastic. I read it about a few months ago and I recommend everybody read it, I just love every minute of it. One issue that I've always thought about comparing the US to the UK is the UK is a settled country, the Industrial Revolution starts way back.

 

It depends on whether it's the early 17th century or before, and grows pretty slow, okay? US is a frontier country and it's really growing very rapidly. And in a sense, it's a dynamic economy. And I've always thought of the business cycles in the US as just a reflection in a sense that it was a newer country.

 

I always thought that banking had a heck of a lot to do with it, but there really is a difference, and I think one way to get at this question and the data is definitely there. You look at the other European countries, there's tons of data for Sweden, Germany, all the Scandinavian countries, France, I mean that might get at that question.

 

 

>> Tyler Goodspeed: I think that's a really good point and-

>> Valarie Ramey: Another way to look at this is to compare to Canada because they didn't have the banking restrictions, but there were also a relatively new country. So then-

>> Tyler Goodspeed: I think that's a really good point. Thank you.

>> Michael Bordo: But I wrote a ton of papers on that with Hugh Rockoff.

 

I have four or five papers comparing US to Canada. So yes, I mean, that completely agrees with what Valerie said.

>> Tyler Goodspeed: And actually, you can see here again with real investment. So this back to 1830 for UK, back to 1901 for US. A lot of volatility in this sort of first half of the 20th century, and for the UK, a little bit earlier.

 

Consumption, I mean, it's interesting just to compare, to keep the vertical axis the same for consumption and real GDP. I mean you just see how much more inertial consumption is over time.

>> John Cochrane: I don't know if I'd call a random walk.

>> Tyler Goodspeed: Yes, fair point. And here again, I mean this is just asking, okay, is there an effect of time on the magnitude of peak to trough declines in GDP and consumption and investment and the unemployment rate?

 

And there is none in basically either economy. If anything post-war period in the United Kingdom that you have some bigger declines, but that's in part because you have that deep recession '43, '47. You have COVID, basically each of the UK's five recessions since the Great Depression were big ones, those are big shocks.

 

I'll try to rush through this because I'm conscious of time and don't worry, I will zoom in on this. But I think just that this perspective is really helpful. The other aspect of recessions that we were perpetually in search of because we like a redemption story or we like to feel that an ordeal was for something.

 

And so we are perpetually looking for cleansing or reallocated or gales of creative destruction aspects of recessions. And don't get me wrong, I mean, and this is not my research, it turns out that yes, creative destruction is super important for growth over the long term, but it is on net not enhanced by recessions.

 

I should just let Bob speak in terms of the productivity ladder that workers climb is generally pro-cyclical. That recovery is about moving from lower productivity to higher productivity firms, and that ladder collapses during recessions. And also, research and development is pro-cyclical. I use a Hamilton trend here because Jim Hamilton tells us that we should never use a Hodrick-Prescott filter.

 

But I also very often use a Hodrick-Prescott filter cuz-

>> John Taylor: Valerie has something to say.

>> Valarie Ramey: Yeah, do not use his method to create a trend.

>> Tyler Goodspeed: Okay.

>> Valarie Ramey: Because see what it does to World War II.

>> Tyler Goodspeed: Yeah, okay, that's really helpful.

>> Michael Boskin: Whose method?

>> Valarie Ramey: Jim Hamilton's method.

 

 

>> Michael Boskin: Don't use it?

>> Tyler Goodspeed: So I do both-

>> Valarie Ramey: By the way, Bob Prescott has a very good working paper that for 2020 discussing the HP versus the Hamilton.

>> Tyler Goodspeed: So I'm agnostic and so I use both. This report just reports the Hamilton, but this is the Median and maximum deviation from the trend share of GDP at the peak of an expansion, the trough of the subsequent recession and the peak of the subsequent recovery.

 

And I mean these are very, very, and it's an absolute because if you, if you don't, if you do mean then you know, different from recession to recession might be average out to zero. So they're all generally very small. And what's interesting is which these bigger ones correspond to.

 

So, oops, this COVID, this is early 1980s when you have sky high interest rates. And actually so it's negative going into the '81 recession. Below trend going into the recession ended even more below trend during the recession. And this is also the early 80s when it was above trend before and above trend at the trough of the recession because of mining structures, so particularly oil and gas.

 

So there was a lot of drilling in response to high oil prices.

>> Robert Hall: Can I ask a question about, I was just looking up some numbers, in 1941 the British population was 48 million while in America was 130. Today it's 68 million versus 330. So great, UK is like a 40% increase in population and the USA is 150% or four times as much.

 

Okay, well doesn't labor go into growth. Now will that have any effect on this kinda stuff?

>> Tyler Goodspeed: I mean, yeah, it'll affect the growth rate, the sort of trend growth rate and level today. But it shouldn't affect sort of the peak to trough fluctuations in that growth would be my intuition.

 

Well actually it could be smaller, could mean smaller peak to trough declines because,

>> Robert Hall: Because you're pouring in people all the time, so you're adding to the labor force.

>> Tyler Goodspeed: Yes, but I mean that.

>> Robert Hall: Unless they're coming in clusters. But you know it's been pretty smooth since the end of World War II, isn't it?

 

I mean we don't know.

>> Tyler Goodspeed: Immigration.

>> John Cochrane: Yeah, I think between growth and fluctuation.

>> Robert Hall: Immigration peaks before 1920. Sort of 1890 to 1910 is a huge peak, and then of course it increases recently.

>> Michael Boskin: And the overall effects about the population are trivial to say.

>> Robert Hall: Yeah, fertility rates are much higher.

 

 

>> Michael Boskin: There's 4 million people added to the population at birth. And that's true whether you look at 1960 or 1980 or even recently, it's always 4 million.

>> Robert Hall: Demography is extremely easy.

>> John Taylor: Literally, that's true

>> Michael Boskin: Yeah, yeah, read the numbers.

>> Tyler Goodspeed: Okay, and actually you can look also at employment by industry sector, similar story.

 

And incidentally here you've got COVID and Valerie might appreciate this. This is a big increase of rise above trend at the end of World War II, followed by a big decline below trend in the immediate aftermath of World War II. The others are all, I mean, the median deviation is 0.1% for all the other categories.

 

Similar story for the UK going back to 1920. Going back before 1920, I can't calculate shares because these are all index values. But we can ask what is the effect of a trend deviation at the peak of an expansion on the trend deviation at the subsequent recession trough?

 

No relation for industry and services, if anything, a positive deviation from trend for agriculture is followed by shallower contraction in the subsequent year or at the subsequent trough. I think that's probably because if you have a bumper harvest, you're probably gonna have some good coal piles and seed stocks for the next year.

 

And you can go into a little bit more detail at the industry level going back to 1870 similar. Now, sorry, I'm rushing through cuz I'm conscious of a time. Now we kinda go back to Hartley's question, which is how correlated are US and UK recessions? Cuz we tend to think of recessions as globalists, but it turns out they're often pretty patriotic.

 

So on a quarterly basis these are the correlation coefficients and the start of a recession. On a quarterly basis, they are basically uncorrelated. And just being in a recession, it's a pretty low correlation. On an annual basis it's a bit higher. And that's been rising over time. So it's about 0.5 the correlation coefficient.

 

But as I noted, the US is just historically much more recession prone and much more likely to lead than lag entry into recession. So since 1700, US has had 68 recessions. Since the Great Depression we've had 15. Since 1700, the UK has had 60 recessions and since the Great Depression they've had five.

 

And I think this differential recessionary experience helps to illuminate what causes recessions. So 2001 is a little bit different. I mean, so the UK sailed through 2001, they were adding 20,000 jobs a month in a much smaller population of costs economy. GDP grew throughout with the US and they didn't have the following shocks that the US experienced that I would list in reverse order of importance.

 

And John, you're gonna shake your head at one of them, but. So no direct 9/11 shock. Second, no permanent normal trade relations shock, which had a very immediate effect on gross hiring in durable goods manufacturing. And this was the first recession for which we had jolts or much more limited energy price shock because by that point UK was relying on coal and North Sea gas.

 

And Valerie and I were talking about this earlier. Gas is a very fragmented market. That is, I mean, cuz it is so much cheaper to ship through pipes. And so those are very fragmented markets. And then the fourth shock the UK didn't have was the bursting tech. I hesitate to use the B word, but a bursting dot com.

 

 

>> Michael Boskin: This observation lies something that's kind of an undercurrent of a lot of the conversation here. UK was more stable, but it's growing at a much lower rate.

>> Valarie Ramey: Exactly, yeah.

>> Michael Boskin: That's the kind of most fundamental question in macroeconomic, right, how much should we worry about recessions?

 

And with an alternative universe where we did a lot more to try to stabilize things, result in much less prosperity over time.

>> Tyler Goodspeed: So I mean, my conclusion is that we worry a lot about recessions and we punish. And actually it's hilarious because we always punish Irving Fisher, brilliant guy, we punish him for that remark about the permanently high plateau.

 

He had some very technical, very academic reasons for saying investment trusts were diversifying, the dollar depreciated so US equity valuations would be higher. He noted a number of other things. That quote was printed in response to Babson, who was another forecaster at the time, who just the day before had given a speech for the National Business Conference.

 

And he said, I repeat what I said at this time last year and the year before that, sooner or later, sooner or later a crash is coming. It goes on sooner or later the stock market boom will collapse like the Florida real estate boom. Someday, the time is coming when the market will begin to slide off sooner or later, someday.

 

He was the 1929 Nouriel Roubini, but he is now permanently celebrated for correctly calling that. And Irving Fisher is denigrated, but,

>> Robert Hall: Paul Samuelson was, he said if you have to forecast, forecast often.

>> Tyler Goodspeed: By the way, yeah, the absence reason when he predicted recession the prior year at the same event was because he said if the Democratic candidate for president won, then we would get a recession.

 

So he just changed the same prediction, but just different rationale. But yes, a fundamental conclusion of the book is we worry a lot about the over the long run, 20% of years in which the economy is in recession. But the US is much more recession prone historically than the United Kingdom is.

 

But we're also 30 to 40% richer per capita.

>> Speaker 14: You look at marginal tax rates.

>> Tyler Goodspeed: I mean, you have all these things that feed into the growth.

>> Speaker 14: Yeah, well, I know, but I think they're very important. And I didn't hear you utter one word, so. Okay, never mind.

 

Just my peek.

>> Michael Bordo: Okay, I wanna get back to the debate you had with Hartley about the transmission of recessions. I mean, there was a ton of work done at the NBER. Anna Schwartz and Lothian and Darby had a big book I did stuff on this. Okay, that under the gold standard, there was a lot tighter correlation between recessions in the US and the UK than afterwards.

 

And you didn't talk at all about the international monetary regime here. And I think that's important.

>> Tyler Goodspeed: Yeah, that is a really good, important point, Mike. And one of the interesting things is drawing inspiration from Friedman and Schwartz's unappreciated book on sort of velocities in relations, interrelations between relations and velocities between the US and UK I think it came out in 1980.

 

I sort of updated that and looked at all these macroeconomic correlations between a bunch of variables and the UK historically imported a lot of nominal shocks from the US but what's really interesting is, I mean, some really high correlations in some of these nominal variables. But look down here, their monetary base isn't changing much.

 

Real consumption is not changing much, the real variables aren't changing much. And the population of banks is the least correlated. So they import a lot of these shocks just on a quarterly basis. But the money supply doesn't change much and the number of banks doesn't change much. And I think this goes back to Hartley's question.

 

I mean these are the number of bank failures. I mean you can't even see the UK bank failures here, see the rate a little bit here.

>> Michael Boskin: There's a lot else besides the rate of failures. First of all, we started with 12,000 banks back then, we have maybe 5,000 now.

 

England has 90% of assets are held by what, five banks maybe something like that. And in Europe traditionally investment banks have been inside large commercial banks. We've also seen this big change. I don't know what the data, hey, you go back to the 80s and early 90s and 70% of credit would be extended by financial institutions.

 

Only 30% direct credit markets, that's flipped and now mostly it's extended direct credit markets. So banking in that sense has become normal banking itself, shadow bag has become much less important. So I don't know what the same thing's going on in the UK but those are gonna compare to the financial system.

 

Those are important things.

>> Tyler Goodspeed: Yeah, and so with the UK, I mean for much of the 18th century they had this six partner rule. Bank can't have more than six partners. So basically, the only way to provide more banking services to a growing British economy was to have more banks, more little banks.

 

So that's why the population per bank was declining, you had a lot of British banking crises during this time, in particular 1825. After 1825 they scrapped that six partner rule and they immediately start to amalgamate to merge. And the most common form of exit from, from UK banking is amalgamation or merger and the population goes up.

 

Whereas in the US I mean I think it troughs at like 4,000 people per bank, like 3, 900 people per bank in 1920, 21. So whereas in the UK, I mean look, you're up at 250,000. I don't think there was actually a drop. It's just different measurements like co ops and stuff.

 

But yeah, that, that is a story of British banking stability because yeah, after 1829, by the end of the 20th, 19th century in the UK you have a few hundred banks. They're all mostly nationally diversified and so much better able to absorb shocks and coordinate responses to shocks.

 

And that's why they ride through 1866. There's a banking crisis, there's no recession, okay.

>> Michael Boskin: What role did the Bank of England play in the year before we had a Federal Reserve, we had the first bank of the United States?

>> Tyler Goodspeed: Typically a coordinating role. So like in 1866 they said to a consortium of private banks, all right, you guys need to mobilize this fund.

 

We'll, we'll top it up, but you guys need to go first.

>> Michael Boskin: So J.P Morgan did that in 1906 or seven?

>> Tyler Goodspeed: Yes, yeah.

>> Michael Boskin: Not finding a huge difference on the founding of the Federal Reserve is kind of a sad application.

>> John Cochrane: Yeah, well, we can talk about that one.

 

We had a long period where we didn't have central bank, put it that way.

>> Speaker 15: Yeah.

>> John Cochrane: But argue what Tyler doesn't show a huge decline in.

>> Michael Boskin: If you're examining the financial system and what's going on, that's a pretty clear difference. What they were actually doing rather than they were just chartered.

 

 

>> Tyler Goodspeed: Okay, and I know Mike has written a ton on this, but as you just said, there was a cost to the UK. So yes, they were super stable. They had a lot of people per bank, a nationally diversified banking system. But that meant that you had a few hundred banks that were not in the business of extending credit to people without much credit history, without much physical collateral.

 

And in 1930 there was a parliamentary commission that, the Macmillan report that identified the Macmillan gap, that was basically UK banks are doing a terrible job of channeling savings into, invest into small and medium sized British enterprises. It was another McMillan, yeah, so yeah, that's a big problem or trade off, you could say for the UK.

 

I'm conscious of time, so I'll just add 1825 to 1947. You could say one of the key reasons for British banking stability was that they didn't have these adverse banking shocks or the amplification of adverse shocks through the banking system. The other period of greater British stability was 1947 to 1973.

 

So they go that entire period without a recession. We had 1948, '53, '57, '60, '70 and then we get to '73. And what's really interesting about those post war, immediate post war recessions in the United States is you don't have to buy Jim Hamilton 100% to recognize that oil supply shocks were a major factor in each of those recessions.

 

And that was at very minimal impact or relatively little impact on the United Kingdom. Why? Because for the duration of that period, coal supplied between 75 and 98% of UK power gen. And you cannot in the short term, those are not substitutes. I mean different grades of oil are not substitutes.

 

But coal and crude oil are not substitutes. You cannot fill a ship, run a ship that was designed to burn coal, to burn oil. You cannot just put oil into a power generation plant. And households, they burn coal, they can't just substitute oil. So that was a really, really big factor.

 

I think in those that 1947 to 1973 period and crucially the UK went that entire period actually they went from 1926 to 1972 without a coal strike. If you go back before then, you have coal strikes in for the 1893 UK recession. 1900, 1902, 1920, '21, '26, '27.

 

And I'm listening both US and UK, and in my two minutes remaining I will take us back to Long and Plosser. I think you have two different types of shocks. You can either have aggregate shocks that affect all sectors relatively indiscriminately even if linkages between sectors aren't particularly high and the elasticity of substitution between inputs is high.

 

Or you can have sector specific shocks in the context of high linkages between sectors and low elasticity substitution at least over a 12 month horizon, which is the duration of the average recession. And both matter. I mean we had credit controls in 1980, the reauthorization of that authority in 1969 we had credit controls in 1949, we had rationing in 1946, '47.

 

Those macro shocks matter, but much more frequent, you go back over the long run are those sector-specific shocks. Energy is a big one in the 18th century, of course an employee of an energy company would say that. But 18th century turf and provender, the biggest thing they mentioned in the aftermath of harvest shocks, it's not the hunger, sorry, it's the famine in fuel, turf and provender, the animal feed for draft power.

 

And then you would often get 1900 1902, 1926. They talk about the fuel famine in coal as a result of the coal strikes. And then you often get in 1857 and 1873. People write about these banking crises as if capital, the nation's capital has gone on strike. The nation's capital will is only willing to work forever higher rates of interest or not at all because they just can't figure out how valuable is the collateral.

 

So I think Long and Plosser we're on to something that those sector-specific shocks are really important. Thank you all.