Susceptibility-based growth predictions

Modern macroeconomic theories were unable to foresee the last Great Recession and could neither predict its prolonged duration nor the recovery rate. They are based on supply−demand equilibria that do not exist during recessionary shocks. Here we focus on resilience as a nonequilibrium property of networked production systems and develop a linear response theory for input−output economics. By calibrating the framework to data from 56 industrial sectors in 43 countries between 2000 and 2014, we find that the susceptibility of individual industrial sectors to economic shocks varies greatly across countries, sectors, and time. We show that susceptibility-based growth predictions that take sector- and country-specific recovery into account, outperform—by far—standard econometric models. Our results are analytically rigorous, empirically testable, and flexible enough to address policy-relevant scenarios. We illustrate the latter by estimating the impact of recently imposed tariffs on US imports (steel and aluminum) on specific sectors across European countries.

That is the abstract of a new piece by Peter Klimek, Sebastian Poledna, and Stefan Thurner in Nature Communications.  Via the excellent Charles Klingman.

Comments

Or maybe modern macroeconomic theories didn’t include a meaningful (non-trivial) banking sector?

What is this non-economics garbage article? Yes, it is obviously garbage because it already accuses "modern macroeconomic theories" from something it never claimed to (which is to predict recessions) in the first sentence. The second sentence claims that supply and demand do not reach equilibrium during a recession, even though experimental evidence on markets (such as the market experiment available Mankiw's textbook which subjects undergrads) is quite clear that they approximate competitive equilibrium after a few rounds of trade. Of course, the authors are not economists such ignorant claims are expected.

Articles like this show that the main problem with economics is not in the interior of the science itself but in the fact that the rest of the world misinterprets what economics is trying to do and show lack of proper humility regarding the obvious fact that they are ignorant of economics and instead act as if their ignorance was some sort of advantage.

Nature has a habit of publishing bizarre ideas from non-economists and framing them as insightful. Another example, https://www.nature.com/articles/s41567-018-0204-y

That is kind of a strange response, unless you are buried deep in what economics wants to be (a paper generating career track?), more than what others want from it (useful advice for life's difficult situations?).

Worried about the next recession? STFU, that's not economics!

Agreed, completely. The authors are not economists and Nature is not peer-reviewed by serious economists; no one in the profession reads this.

If only no one would read what serious economists write, isolation would be complete.

The best way for economics to stay relevant is to predict stuff. If you refuse to predict, don't be surprised when someone else takes up the challenge.

You are not quite right in this point: two of the three authors are not only physicists and complexity scientists but also economists.

I instinctively tune out whenever the inability to predict an extreme event is used as a criticism of economic theory; it is an exceedingly bad criticism.

The inability to explain the response, however, is fair game.

Business cycles might creste the single biggest impact "economics" makes on human lives. They are certainly the most sudden, once you have developmental economics and a general growth path squared away.

That's why I am pushing back on this "how dare you ask us for a warning bell" stuff.

It's a fair request from Main Street.

And I’m sure you are leading the marches down Main St heeding the warning about debt service, Social Security and Medicare payments that are unsustainable. I’m also certain that as someone who self-describes as center-right is telling those on Main St to contact their representatives and senators and call for immediate cuts in these programs. Especially since you believe that Democracy requires a watchful eye.

Eck-shually, we center-pragmatists do have a lot to complain about with (1) the economic imbecility of the Trump administration, and (2) the conspiracy of silence from the economic profession on the same.

It's like, during the Obama administration, blogs like MR were all about critiquing economic strategies. Which they should. That's one of their jobs.

So what was their response in response to an administration that flat wasn't going to listen to good advice?

To return to the roots of economics, that is papers no one reads.

Because to critique would be to admit the deafness.

It's not a "fair request" because it has no epistemological justification. It is fundamentally impossible to predict extreme shocks -- that's why they're called extreme shocks. Just because you would like to be able to predict them doesn't mean it's sensible to criticize economists for not being prophets.

It is just as unreasonable to expect a seismologist to predict an earthquake. The sensible question to ask is why so many buildings were not up to code when the earthquake hit, and why it took so long to rebuild, to stretch the analogy.

I'll wait patiently while the authors' methods fail to foresee the next recession and fail to predict its duration and recovery rate, i.e. the very inabilities they cite that afflict "modern macroeconomic theories."

Life for economists was easier when they could rely on predictable economic cycles for booms and busts. Here's an article from last week in the NYT in which the author is obviously frustrated by the absence of a consensus for what lies ahead: https://www.nytimes.com/2019/04/12/business/wall-street-lose-faith-in-fed.html One will note that the issue is framed from the perspective of Wall Street (i.e., whether asset prices will go up or go down). A problem with reliance on rising asset prices for prosperity is that the mind of the investor is hard to predict (especially when compared to more predictable economic cycles).

Years ago, I lost faith in the Fed. It was established in 1913 (same year as the income tax - coincidence?) to abolish/smooth over business cycles. It has provided an unbroken record of failure from day-one.

The original Fed had four powers/responsibilities. The most vital was lender of last resort to provide liquidity (a concept far above economists' pay grades) to member banks in cash flow peril by discounting credit-worthy (loan) paper thus providing necessary cash/liquidity until conditions improve. Now, after failing to abolish depressions/recessions, idiots in Congress [redundant] gave the Fed power to monkey with employment, money supply, rates, prices/inflation, etc. Go figure.

Agree that rising prices are not real GDP growth/prosperity. The more money the Fed pushed into the economy the worse the economic conditions became. More money weakens wealth generation by stimulating consumption that is not preceded by the production of real wealth. We big-time saw this in the run-up to the 2008 financial catastrophe when housing prices skyrocketed - too much money chasing housing - while real disposable incomes, GDP growth etc. were comparatively flat.

Here is one "take" on recent business cycle: Boom/Bust - based on speculation, get-rich-quick schemes, and over-supplies of cash from easy/low-cost credit. In 1873 it was speculation in RR's. In 1929 it was massive speculation in stocks on margin. In 2006 it was over-leveraged residential RE and mortgage derivatives (more a symptom). Common among all:: bank failures/runs, chaos, asset prices fall, stocks crash, panic. Fed reaction - add liquidity. Before the Fed (since the CW National Banking Act) the Treasury could purchase T bonds from N/B's and reissue greenbacks.

Here is our host on the benefits of the deflation following the 1873-79 depression: https://marginalrevolution.com/marginalrevolution/2011/08/the-deflation-of-1873-1896.html I don't recall a similar blog post about the deflation following the 1929 financial crisis and great depression, but the logic to support (not oppose) it would be the same. And so would the logic for opposition to the Fed's reaction to the 2008 financial crisis and great recession (i.e., to reinflate assets). I might point out that the economic conditions preceding these economic crises were the same: a high level of inequality, inequality produced in large part by rising asset prices. Of course, Cowen's blog post about the 1873-79 deflation is right out of the Austrian's playbook.

With the admittedly huge exception of the Great Depression it's hard to see the 20th century US economy as more volatile than what came before. That realtive smoothness may or may not be the Fed's doing but it certainly isn't evidence that the Fed has failed.

You missed the point: deflation was a good thing not a bad thing (according to the Austrians). Read Peter Boettke on his advice during the great recession: let asset prices fall! Boettke can be admired for his honesty if not for his propensity for Kinsley gaffes.

Deflation is only a good thing for the owning class. For everyone else it sucks. Moreover a small level of inflation is natural, rather like entropy in physics. You can tamp it down, maybe, and locally gor short periods of time, reverse it, but over time accumulated stocks of wealth will erode in value much as black holes do via Hawking Radiation. People should plan their affairs accordingly.

It seems like this kind of statistical work would lead itself to the use of AI analysis. Current predictive software is reasonably good at taking many data inputs over a series of time and predicting future response.

I assume someone has done this type of work.

No. Typical AI applications require millions or billions of data points to estimate a complicated model (speech, handwriting, photo recognition). We only have a few years of macroeconomic data, and can only estimate simple stylized models.

The paper just estimates a linear matrix model and compares it to a univariate model. But we already had Litterman's vector-autoregression forecasting in the 1970's. Moreover, these simple statistical models do not have causality, and are useless for policy analysis (Lucas Critique). Economics is not largely about forecasting.

President Captain Bolsonaro has oficially been chosen by famous American magazine Time as one of the most influential people in the world. He has also been chosen by the Brazilian-American Chamber of Commerce as Person of the Year. Meanwhile, his rival, corrupt leftist Mr. Haddad, has been sent to the dustbin of History. I am tired of winning.

Wow! You mean, like, if you disaggregate a nationwide economy into 56 industries you get less noise than you do with aggregate nationwide macroeconomic data? Who would have thunk it?

Somewhere/when a 70s Soviet planner want's their model back.

Connecting all our balance sheets in a network of correlation links, suitably trimmed. Not a bad idea. It is not symmetric flow, however. It is less I more O, to cover uncertainty.

With so many parameters, I can fit whatever system.

Folks, economists have been borrowing from physics for over a hundred years. Fischer Black, who is so popular at GMU and in finance departments, trained as a physicist.

It is not a coincidence that so much pioneering economics since 1945 came from MIT, where the economists borrow techniques from the hard sciences.

The use of calculus and optimization theory, continuous time dynamics, stochastic calculus, statistical theory: all borrowed from physics.

Maybe there is a lot more analytical technique that economists can borrow. There is always chaos theory. And physics have studied "frictions" for a long time.

There is a saying that at Chicago, macro is just micro with money.

Maybe macro is really just micro with frictions. The reason for the failure of the Arrow-Debreu model or neoclassical economics is the absence of frictions.

"Modern macroeconomic theories were unable to foresee the last Great Recession and could neither predict its prolonged duration nor the recovery rate."

The Austrian School guys were certainly predicting a major decline in housing that would create a significant slowdown for the economy. Many pointed out that you cannot predict how long the contraction would last because that depends on the success that the interventionists at Treasury and the Fed have at hiding the fact that they are corrupt charlatans who have no real incentive in proposing a meaningful reform of the system.

The fact that the mainstream economic industry is broken is not a surprise. The surprise is how long they have fooled the public.

I love reading the comments. I always learn so much.

Comments for this post are closed