Category: Data Source

Measuring the Cost of Regulation: A Text-Based Approach

We derive a measure of firm-level regulatory costs from the text of corporate earnings calls. We then use this measure to study the effect of regulation on companies’ operating fundamentals and cost of capital. We find that higher regulatory cost results in slower sales growth, an effect which is mitigated for large firms. Furthermore, we find a one-standard deviation increase in our preferred measure of regulatory cost is associated with an increase in firms’ cost of capital of close to 3% per year. These findings suggest that regulatory risk is a major cost to firms, but the largest firms are able to manage that risk better.

That is the abstract of a new NBER paper by Charles W. Calomiris, Harry Mamaysky, Ruoke Yang, a piece written in pre-Covid-19 times.  It has never been more relevant, except that the estimates for regulatory costs turn out to be far too low (no criticism of the authors is intended here).  To repeat my earlier point, America’s regulatory state is failing us.

That was then, this is now: Palantir privacy edition

Data-analytics company Palantir Technologies Inc. is in talks to provide software to governments across Europe to battle the spread of Covid-19 and make strained health-care systems more efficient, a person familiar with the matter said.

The software company is in discussions with authorities in France, Germany, Austria and Switzerland, the person said, asking not to be identified because the negotiations are private…

European Union Commissioner Thierry Breton said Monday that the bloc is collecting mobile-phone data to help predict epidemic peaks in various member states and help allocate resources.

Palantir has signed a deal with a regional government in Germany, where it already has a 14 million euro ($15 million) contract with law enforcement in North Rhine-Westphalia, the person said. Palantir is also seeking a contract at a national level, the person said, but talks have stalled, the person added.

When a nation or company buys access to Palantir, it can use the data analytics software to pull far-flung digital information into a single repository and mine it for patterns.

Here is the full story.  From a distance it is difficult to evaluate these deals, but I will stick with my general claim that the anti-tech intellectuals have become irrelevant, and for the most part they know it.

The fiscal multiplier during World War II

WWII is viewed as the quintessential example of fiscal stimulus and exerts an outsized influence on fiscal multiplier estimates, but the wartime economy was highly unusual. I use newly-digitized contract data to construct a state-level panel on U.S. spending in WWII. I estimate a relative fiscal multiplier of 0.25, implying an aggregate multiplier of roughly 0.3. Conversion from civilian manufacturing to war production reduced the initial shock to economic activity because war production directly displaced civilian manufacturing. Saving and taxes account for 75% of the income generated by war spending, implying that the add-on effects from increased consumption were minimal.

That is from a 2018 paper by Gillian Brunet, and you will note that it reflects the consensus of the literature as a whole.  I do favor the federal government borrowing and spending a great deal of money right now on things that we need.  If you think we are in a traditional Keynesian scenario, or are pulling out a traditional AS-AD model, you are going to be very badly disappointed.  Most of all, we need to be spending more on public health and remedies for Covid-19.  Here is my earlier Bloomberg column on analogies and disanalogies between Covid-19 and World War II.  And again, see Garett Jones and Dan Rothschild on the 2009 stimulus.

Did unconventional interventions unfreeze the credit market?

By Hui Tong and Shang-Jin Wei, newly relevant!

This paper investigates whether and how unconventional interventions in 2008–2010 unfroze the credit market. We construct a dataset of 198 interventions for 16 countries during 2008–2010 and examine heterogeneous responses in stock prices to the interventions across 7,873 nonfinancial firms in those countries. Stock prices increase when the interventions are announced, particularly for firms with greater intrinsic need for external capital. This pattern is corroborated by subsequent expansions in firm investment, R&D expenditure, and employment. Among various forms of interventions, recapitalization of banks appears particularly effective in channeling the intervention effects from financial to nonfinancial sectors.

That is from the new issue of AEJ: Macroeconomics.

A new NBER paper on the Wuhan lockdown

We quantify the causal impact of human mobility restrictions, particularly the lockdown of the city of Wuhan on January 23, 2020, on the containment and delay of the spread of the Novel Coronavirus (2019-nCoV). We employ a set of difference-in-differences (DID) estimations to disentangle the lockdown effect on human mobility reductions from other confounding effects including panic effect, virus effect, and the Spring Festival effect. We find that the lockdown of Wuhan reduced inflow into Wuhan by 76.64%, outflows from Wuhan by 56.35%, and within-Wuhan movements by 54.15%. We also estimate the dynamic effects of up to 22 lagged population inflows from Wuhan and other Hubei cities, the epicenter of the 2019-nCoV outbreak, on the destination cities’ new infection cases. We find, using simulations with these estimates, that the lockdown of the city of Wuhan on January 23, 2020 contributed significantly to reducing the total infection cases outside of Wuhan, even with the social distancing measures later imposed by other cities. We find that the COVID-19 cases would be 64.81% higher in the 347 Chinese cities outside Hubei province, and 52.64% higher in the 16 non-Wuhan cities inside Hubei, in the counterfactual world in which the city of Wuhan were not locked down from January 23, 2020. We also find that there were substantial undocumented infection cases in the early days of the 2019-nCoV outbreak in Wuhan and other cities of Hubei province, but over time, the gap between the officially reported cases and our estimated “actual” cases narrows significantly. We also find evidence that enhanced social distancing policies in the 63 Chinese cities outside Hubei province are effective in reducing the impact of population inflows from the epicenter cities in Hubei province on the spread of 2019-nCoV virus in the destination cities elsewhere.

That is by Hanming Fang, Long Wang, and Yang Yang, of course do beware data quality issues.

Where does all the heterogeneity come from?

Here is a Christopher Balding tweet storm, excerpt:

Iceland has done almost 14k tests on an island of 360k so more than 3% of the total population…They have more than 800 confirmed cases, 10k people in quarantine, 800 in isolation, 18 hospitalizations, 6 in ICU, and 2 dead…About how many people SHOULD have corona if the spread etc numbers are accurate. As of March 27, Iceland would be expected to have more than 46k people that have corona. Emphasis this is on an island of 360k and 800 confirmed cases.

What is going on in the Icelandic numbers?  What accounts for this apparent heterogeneity?  Dosage?  Is it that Icelandic clustering is mostly in one easy to control central city and the rest already is “socially distanced,” even in the best of times?

I know there are some MR readers in Iceland, and presumably they read the Icelandic press.  Can anyone shed light on why the death rate is not higher in Iceland?  Is it that the death rate is about to burst a week from now?  Alternatively, you might think the Icelanders have kept their hospitals up and running — important for sure — but that doesn’t explain what seems to be a quite low rate of reported cases.  Or is it that Iceland’s second largest city is so tiny — Akureyri at 18,925 inhabitants — that the virus doesn’t have many easy chances to recirculate once cut off for a while?

Similarly, Sweden hasn’t restricted public life very much and they do not seem to be falling apart?

How much better is Staten Island (less dense) doing than Manhattan (more dense)?

Some reports indicate that in hard-hit Westchester County,. NY, the rate of hospitalization is about one percent (8-10 percent in some other places).  Alternatively, here is serious talk that the death toll in Wuhan is 20x official figures.

How much of the heterogeneity results from the kind of mixing you get?  One account of the low German death rate is the young and the old were never pushed together so much by the policy response.  One account of the high Italian death and hospitalization rate is that the initial quarantine was only regional and thus it spread very dangerous forms of mixing throughout the larger country.

It is possible that Cambodia, Thailand, and Vietnam still will be hit hard, but so far the signs do not indicate as such.  Warm weather may play a positive role, though that remains speculative.  The latest weather paper appears credible and indicates some modestly positive results.  Of course weather won’t explain the relative Icelandic and Swedish success, if indeed those are truly successes.

By the way, on the “everyone already has it” theory, a semi-random sample of 645 from Colorado showed zero positives.

So where is all this heterogeneity coming from?  Is it all just bad data?  That seems hard to believe at this point, and Iceland seems like a plausible source of reasonably good data.

As for concrete conclusions, these heterogeneities should make us more skeptical about any models of the situation.  But it would be wrong to conclude that we should do less, arguably risk-aversion could induce us to wish to do more, including on the lock downs front.

It is also worth pondering which heterogeneities are “baked in,” such as heat and age structure of the population, and which heterogeneities can be altered at the margin, such as forms of social mingling.  It is at least possible that studying these heterogeneities could make policy far more potent.

Overall, I do not see enough people asking these questions.

Are many more people infected than we think?

Here is the Clive Cookson FT piece (with an irresponsible headline).  Here is the Lourenco new Oxford study, only a few pp.  Miles Kimball offers analysis and numerous references.

Here is the Bendavid and Bhattacharya WSJ piece that perhaps has had the biggest popular influence.  They argue that many more people have had Covid-19 than we think, the number of asymptotic cases is very large, and the fatality of the virus is much lower than we think, perhaps not much worse than the flu.  But their required rate of asymptomatic cases is implausibly high.

The best evidence (FT) for asymptomatic cases ranges from 8 to 59 percent, and that is based on a number of samples from China and Italy, albeit imperfect ones.  Icelandic data — they are trying to sample a significant percentage of their population — suggest an asymptomatic rate of about 50 percent.  To be clear, none of those results are conclusive and they all might be wrong.  (And we should work much harder on producing better data.)  But so far there is no particular reason to think those estimates are wrong, other than general uncertainty.  You would have to argue that the asymptomatic cases usually test as negative, and while that is possible again there is no particular reason to expect that.  It should not be your default view.

Marc Lipitsch put it bluntly:

The idea that covid is less severe than seasonal flu is inconsistent with data and with the fact that an epidemic just gathering steam can overwhelm ICU capacity in a rich country like Italy or China.

Furthermore, the “optimistic” view implies a much faster spread for Covid-19 than would fit our data from previous viral episodes, which tend to come in waves and do not usually infect so many people so quickly.

So I give this scenario of a very low fatality rate some chance of being true, but again you ought not to believe it.  The positive evidence for it isn’t that strong, and you have to believe a very specific and indeed unverified claim about the asymptomatic cases testing negative, and also about current spread being unprecedentedly rapid.

Here is Tim Harford’s take (FT) on all this, he and I more or less agree.

By the way, Neil Ferguson didn’t walk back his predictions.  That was fake news.

So we still need to be acting with the presumption that the relatively pessimistic account of the risks is indeed true.  Subject to revision, as always.

The coronavirus situation in Japan is probably much worse than you think

I have been corresponding with a working group regarding the covid-19 situation in Japan.  They shared a draft of their white paper with me while attempting to circulate their revisionist conclusions in policy circles.

The speed premium is indeed increasing quickly.  The white paper has not materially changed since when I first saw it. Since then, the Olympics were postponed and experts in Japan have described the outbreak as “rampant.”  The working group feels that society needs to prepare, and that this outweighs the desire to wait for additional official confirmation.

The authors are an international team based in Tokyo. They cannot attach their identities to the white paper at present.  They are not medical researchers. They have reviewed their conclusions with a medical researcher and others.  You can weigh the evidence of their claims.

Here is the document (no, it is not malware), and here is the opening bit:

The governmental and media consensus is that Japan is weathering covid-19 well. This consensus is wrong. Japan’s true count of covid-19 cases is understated. It may be understated by a factor of 5X or more. Japan is likely seeing transmission rates similar to that experienced in peer nations, not the rates implied by the published infection counts. The cluster containment strategy has already failed. Japan is not presently materially intervening at a social level. Accordingly, Japan will face a national-scale public health crisis within a month, absent immediate and aggressive policy interventions.

There is a great deal of further detail, including the numbers, at the link.  Sobering.

Lessons from the “Spanish Flu” for the Coronavirus’s Potential Effects on Mortality and Economic Activity

That is the subtitle of a new paper by Robert J. Barro, José F. Ursúa,  and Joanna Weng, here is the abstract:

Mortality and economic contraction during the 1918-1920 Great Influenza Pandemic provide plausible upper bounds for outcomes under the coronavirus (COVID-19). Data for 43 countries imply flu-related deaths in 1918-1920 of 39 million, 2.0 percent of world population, implying 150 million deaths when applied to current population. Regressions with annual information on flu deaths 1918-1920 and war deaths during WWI imply flu-generated economic declines for GDP and consumption in the typical country of 6 and 8 percent, respectively. There is also some evidence that higher flu death rates decreased realized real returns on stocks and, especially, on short-term government bills.

I wonder if the economic cost isn’t higher today because we know more about how to limit pandemic spread and we also value human lives more, relative to economic output?

Kudos to the authors for such swift work.

Also from NBER here is Andrew Atkeson on the dynamics of disease progression, depending on the percentage of the population with the disease.  Here is an excerpt from the paper:

Even under severe social distancing scenarios, it is likely that the health system will be overwhelmed, which is indicated to happen when the portion of the U.S. population actively infected and suffering from the disease reaches 1% (about 3.3 million current cases).7 More severe mitigation efforts do push the date at which this happens back from 6 months from now to 12 months from now or more, perhaps allowing time to invest heavily in the resources needed to care for the sick. It is clear that to avoid a health care catastrophe as is currently being experienced in Italy, prolonged severe social distancing measures will need to be combined with a massive investment in health care capacity.

Under almost all of the scenarios considered, at the peak of the disease progression, between 10% and 20% of the population (33 – 66 million people) suffers from an active infection at the same time.

A not entirely cheery prognosis.

Pooling to multiply SARS-CoV-2 testing throughput

Here is an email from Kevin Patrick Mahaffey, and I would like to hear your views on whether this makes sense:

One question I don’t hear being asked: Can we use pooling to repeatedly test the entire labor force at low cost with limited SARS-CoV-2 testing supplies?

Pooling is a technique used elsewhere in pathogen detection where multiple samples (e.g. nasal swabs) are combined (perhaps after the RNA extraction step of RT-qPCR) and run as one assay. A negative result confirms no infection of the entire pool, but a positive result indicates “one or more of the pool is infected.” If this is the case, then each individual in the pool can receive their own test (or, if we’re getting fancy [read: probably too hard to implement in the real world], perform an efficient search of the space using sub-pools).

To me, at least, the key questions seem to be:

– Are current assays sensitive enough to work? Technion researchers report yes in a pool as large as 60.

– Can we align limiting factors in testing cost/velocity with pooled steps? For example, if nasal swabs are the limiting reagent, then pooling doesn’t help; however if PCR primers and probes are limiting it’s great.
– Can we get a regulatory allowance for this? Perhaps the hardest step.

Example (readers, please check my back-of-the-envelope math): If we assume base infection rate of the population is 1%, then pooling of 11 samples has a ~10% chance of coming out positive. If you run all positive pools through individual assays, the expected number of tests per person is 0.196 or a 5.1x multiple on testing throughput (and a 5.1x reduction in cost). This is a big deal.

If we look at this from the view of whole-population biosurveillance after the outbreak period is over and we have a 0.1% base infection rate, pools of 32 samples have an expected number of tests per person at 0.0628 or a 15.9x multiple on throughput/cost reduction.

Putting prices on this, an initial whole-US screen at 1% rate would require about 64M tests. Afterward, performing periodic biosurveillance to find hot spots requires about 21M tests per whole-population screen. At $10/assay (what some folks working on in-field RT-qPCR tests believe marginal cost could be), this is orders of magnitude less expensive than mitigations that deal with a closed economy for any extended period of time.

I’m neither a policy nor medical expert, so perhaps I’m missing something big here. Is there really $20 on the ground or [something something] efficient market?

By the way, Iceland is testing many people and trying to build up representative samples.

Why such a large difference in fatality rates across European nations?

Here is a relevant tweet thread started by Moritz Kuhn, many interesting comments.  For instance Moritz writes: “What is more, it may provide a warning sign for those countries where the elderly and the young live close together, how important it is to contain the virus there early on. These countries are within Europe in particular such as Serbia, Poland Bulgaria, Croatia, or Slovenia.”

Also on Italy, Dan Klein writes to me:

  1. They kiss, hug more, converse longer.
  2. Young people live with their parents, family more.
  3. They smoke somewhat more (packs smoked per capita twice that of Sweden). Smoking weakens the lungs. But also we smokers finger and thumb our cigs and then put them into our mouth. Wash hands first!
  4. For these and whatever adventitious reasons, Italy was early to the problem, and it spread before people learned to adjust behavior.

We will learn more soon.

It matters that more older people are working now

Working women – For the first time, there are now more women aged 60-64 in work than not, analysis of data from the Office for National Statistics shows. The shift has been triggered by changes to the state pension age, the data reveals, with the number of older women in the workforce increasing by 51% since the reforms were introduced in 2010. The number of working men aged between 60 and 64 increased by 13% over the same period. Experts described the shift as “seismic” and said it would have profound implications for women now and in later life.

That is from the Guardian about the UK, via Stephen Gibbons.

The Danes wish to know more about the coronavirus

From Martin Permin, for formatting reasons I have imposed no further indentation:

“Thanks for the excellent coverage on MR.

I lead a small team of tech workers in Copenhagen, who are donating our time and money towards building a covid-19 self-reporting tool for those citizens not (yet) in contact with health care services.

As countries shift from containment to “flatten the curve” strategies, authorities lose track of the number of non-critical cases, and to which degree people adhere to social distancing dictums. This makes it hard to predict the number if ICU beds needed a few days into the future. We’re aiming to solve this by asking all Danes for daily status updates.

Denmark is a good testing ground, but we’ll open source everything, and are building with developing countries in mind. We’re aiming to launch Monday — currently working on a green light from local health authorities.

We’re determining which data to collect. We’d love it if you’d help by asking your audience: “What daily self reported measures would you most like to see from the complete population Denmark?” (or some variation thereof).

There is of course a tradeoff between data fidelity and engagement.

What we’re considering:

  • Symptoms
    • Binary
    • Type
    • Severity
  • Whereabouts
    • Degree of social distancing
    • Hygienic measures
  • Moral
    • How concerned are you
    • Do you know anyone who’s been sick”

Are there comparable efforts to do this elsewhere?