25 Responses to “Take the Tablet”

1. 1 1 Advo
   June 29, 2020 at 6:13 am

   About the farmers (3rd world and medieval England) – I’m guessing information and collective action problems prevent this swap.
   Germany had a similar issue, which as far as I know was only fully fixed after WWII.

   The way I see it, if you have two farmers who can just swap plots of identical value, sure, no problem.

   But in reality, you have perhaps 50 farmers, who mostly have plots of significantly differing size, yield and value.
   To organize a land swap, you need to pool ALL that land (or at least the large majority of it) and have some neutral arbiter reallocate it.
   Unfortunately, you don’t have a neutral arbiter because one way or the other, every local will have a big stake in the outcome of the reallocation process (and if it’s only because his relatives are affected).
   And even if there were such a neutral arbiter, he will work on the basis of incomplete information. He doesn’t know the average long-term yield of a plot of land.

   If this is a voluntary scheme, the farmers who have low-yielding land would be eager to participate, hoping to get a similarly sized, better located and higher-yielding piece of land, whereas the farmers with high-yielding land would rightly expect that they’ll come out worse.

   If you’re struggling at subsistence level, you get really risk-averse.

   Note that this would all be much less of an issue if people had access to capital. Then it could be possible to consolidate at least some of the land via individual transactions between parties. But they don’t.
   As usual, if a population (or a segment of the population) does not have capital or access to capital, the economy will be very inefficient.

2. 2 2 Advo
   June 29, 2020 at 6:28 am

   In addition, you have transaction costs. You’d need to survey all those plots (many of them with quite irregular shapes). The locals will not have the knowledge and/or tools to do so, and they can’t hire a professional surveyor, because even if they knew how to get a hold of a surveyor (which they generally probably don’t), they don’t have the capital and the cost of the surveyor together with any other transaction cost would perhaps even outstrip the benefit of the whole exercise.

3. 3 3 Coupon_clipper
   June 29, 2020 at 8:28 am

   SL, you are an exceptional writer. You covered some advanced topics in your usual lucid manner.

   And usually when people mention politics or individual politicians, the writing just turns into angry yelling that turns off readers. You actually brought up politics towards the end of the piece in a funny, even-handed, and useful way.

4. 4 4 Bill
   June 29, 2020 at 11:59 am

   In ” … the average American income in 1968 was about 40% of what it is today. According to various estimates, this probably meant that the value of a life was only about $6 million back then.”

   Should that read $4 million?

5. 5 5 Steve Landsburg
   June 29, 2020 at 12:17 pm

   Bill: No, this is correct. The income elasticity of the value of life in developed countries is estimated to be about .6, so a 60% reduction in income should yield about a 36% reduction in value of life. From there, I rounded (since the .6 is only a rough estimate in the first place).

6. 6 6 Steve
   June 29, 2020 at 1:58 pm

   I agree with most of it but not sure I can support this part:

   “…you don’t need mandates insofar as you believe the main problem is transmission in stores, bars and restaurants, where owners are fully incentivized to look out for their customers’ interests”

   I think “fully” incentivized is a step too far. They are immensely incentivized to ensure nobody slips and falls in their store and sues them. They are incentivized to avoid bad press from a publicized outbreak of COVID-19 linked to their business. I think the incentives to ensure a virus-free experience is probably clouded somewhat by a free-rider-type situation. One restuarant/bar that chooses the loosest social distancing/mask policy gets the most business (from those who don’t care about the virus but are not immune to it no matter how fake they claim it to be). Other restaurants gain little from doing their part to limit occupancy if others are going to throw caution to the wind and risk another lockdown anyways.

   Thoughts?

7. 7 7 Advo
   June 30, 2020 at 5:17 am

   >>>I think the incentives to ensure a virus-free experience is probably clouded somewhat by a free-rider-type situation<<<

   There's also an externality problem. The main risk from the virus is not to the (young, fit, healthy) customer, but to the (older and/or fatter and/or unhealthy) people he comes into contact with.

   Plus young people don't look at risk the same way as older people. That's why young people make such good infantry.

8. 8 8 Harold
   June 30, 2020 at 5:18 am

   Your back of the envelope numbers are in the region of maybe, maybe not worth it.

   If we spend a lot of money saving the lives, we will have less money and the value of a life will go down. Do we use the old value or the new one? As with most basic economic calculations the individual transactions are assumed not to alter the price, but in this case that might not be correct.

   I do not think you deal with the externalities sufficiently. Reducing social contact has a joint benefit that I do not think can be replicated by individual property owners or individual responses. If R is above 1, the epidemic increases. If it is below 1 it reduces. It does not take very many people to be risk averse for the epidemic to simply keep on growing, and then everyone is at risk. If the epidemic grows too far, the economic costs may follow anyway, but with more deaths.

   The Nordic countries have presented us with something of a natural experiment. Not one that would translate directly to other countries, but they do give us a feel. Sweden as 65,000 cases and 5,300 deaths to date using a more relaxed approach. Norway has 9,000 cases and 250 deaths with a strict lock-down. Finland has 7,000 with 320 deaths. We won’t know which is the better approach until we have the economic data, and future disease data. It does seem to show that the more serious lockdown has reduced the number of deaths, and presumably has resulted in greater economic costs

   Worldometers has projections. They show by September, Norway having -9% social mobility, Sweden at -6% and Finland at -1%. At the lowest, Norway had observed -60% and Sweden -34% and Finland -56%. The current data is suspect, so we must treat these projections with a great deal of caution, but it gives us some point of discussion. The lives saved by the more severe social mobility will be about 5,500 if we assume everything else is the same. What we don’t know is the economic cost of reducing social mobility to 60% rather than 34%.

   The USA got to -54% at the trough. Italy to -79%.

   Link here – well worth a dig around.
   https://covid19.healthdata.org/united-states-of-america

   I suspect it is impossible to get below -30% without some form of compulsion at the start of an epidemic, but that is off the top of my head. Sweden did have some compulsion and restrictions and it got to only -34%. However, that could change if say 20% of the population was infected and one time. Then the chances of catching it would be very high for every trip out, and everyone may choose to stay at home. That would give the worst of all worlds of maximum deaths and maximum economic harm.

   It is at least plausible that compulsory lock-downs result in less damage overall than leaving it to individual choice.

   The HANS neck device is interesting. Drivers were not willing to use the device until it was forced on them, even though it apparently enabled them to drive 2% faster! A strange choice for racing drivers. That is a significant performance improvement and at the top level a 2% performance improvement would make the difference between champion and also ran (or possibly not even running at all).

9. 9 9 Bill
   June 30, 2020 at 9:48 am

   Steve, thanks. I missed the part in your essay regarding the income elasticity of value of life estimate.

10. 10 10 joe henry
    June 30, 2020 at 7:47 pm

    * I agree with Coupon Clipper #3, Steve does have a knack for replacing emotion with, well, uh, Fun!

    * Lifted from the article;

    ” Now let’s do a different experiment with the same urn. Once again you get to draw one ball. This time, would you rather bet against the ball being red or bet against the ball being green?”

    ” Most people choose to bet against red. Apparently they believe that the reds are outnumbered by the greens!”

    ” It’s fine to believe that the reds outnumber the greens, and fine to believe that the greens outnumber the reds. But when you’re able to believe both things at once, I start to worry about how you survive in the world.”

    Uh, no. I think many, me included, simply misread it. It sounded like we had 60 non-red balls (green, purple, black, white, big, small) against 30 red balls. (The game changed.)

    * Now, just for the sake of rambling on (where else can an old man ramble on?) What is the value of life? Yes, I read the article but I’m old. I’m now referred to as a “Useless Eater”. I consume more than I produce. My value is a negative. Even though politicians might “say” I have more value than a cheap hamburger the reality is that they would *Love!* it if I and my ilk would quit withdrawing what we paid into Social Security and Medicare.

    As would the “Green New Deal” gang. The bottom line it is people who contribute to “Global Climate Change”. A mere billion people can do pretty much whatever they like. Seven Billion … not so much.

    COVID was the perfect opportunity for politicians to balance the budget and environmentalists to reduce CO2. Maybe NYS governor Cuomo had it right when he moved COVID patients into nursing homes?

    What happened?

11. 11 11 Harold
    July 1, 2020 at 8:42 am

    #10. The urn example does seem to be phrased in a somewhat confusing manner, but it does make sense.
    Bet 1 – is the ball you draw going to be red or is the ball green? Most bet red. This is straightforward – if you think there are more reds you bet red. If it is neither red nor green I guess you keep your stake? Say for the sake of argument you are working on a model that there are 10 greens.

    Bet 2 – is the ball going to be not red or is the ball going to be not green? If you keep your working model that there are 10 greens, then the chances of it being not green are 80/90. The chances of it being not red are 60/90. Not green should be your bet. Yet most people bet not red.

    That is how I interpret it anyway. It would be consistent to bet not green, yet most people bet not red. Apparently.

    The real problem is that we have no idea how many greens there are. We are really betting on whether there are more reds than greens, and as far as we are aware that is probability 1/2. Either it is or it isn’t and we have no further information. So with a 50/50 bet, we make an arbitrary choice that can be tipped one way or the other for infinitesimal reasons. It is just tiny fluctuations that change our bet, such as how the question is posed.

    Steve has presented much more in-depth ball choosing examples that demonstrate a much stronger change in preference. I assume this was a simplified example for the purposes of brevity.

    The main point is that we behave inconsistently in the lab, in life we are consistent. I am not sure this is actually true, although I think we can say we are more consistent when the stakes are real. For all the examples Steve cites, I suspect we could come up with examples where house prices have been depressed much more than the real risk associated with its location and many similar examples. If we concede that people generally are pretty good at this assessment, I am not sure we can say there are not situations where it goes badly wrong as well.

    This presents a possible problem for Covid-19. We have no experience of this sort of thing. Is it one of these situations where people react inconsistently with the real risk? Part of the problem is we do not even know with any certainty what the real risk is – we certainly did not when the behaviours needed to change to have any effect. Germany locked down when their rate of deaths was 1/4 that when UK locked-down. That number approximately tracked through the outbreak so far. It is possible that they reduced the number of deaths by tens of thousands by 1 extra week or so of lock-down. It also means they can come out of lock-down earlier and faster, thus reducing both direct economic cost and number of immediate deaths.

    It is interesting that the reason the UK was slow was because they relied on the much-maligned Ferguson report. It estimated a doubling time of 5-6 days when it was actually 3 days. Ferguson was “wrong”, but was underplaying rather than exaggerating.

    This response on these timescales is simply not possible by appealing to self assessment of risk.

    The big question – what is the “better” outcome and how would we measure it? We know for a fact that if we had left it to individuals assessment we would have had a much greater initial outbreak and far more dead by now. There is a 2 week or so delay between the infections and seeing the dead really rise. What we don’t know is how people would have reacted to that. Maybe we just sacrifice the older people, as Joe above talks about? Maybe there will be no treatment and no vaccine and these people will all eventually succumb anyway.

12. 12 12 joe henry
    July 1, 2020 at 4:26 pm

    Harold #11; Thank you for your time and your thoughtful response.

    In game 1 I have a one in three chance of being red against some unknown. It’s a crap shoot. Flip a coin. I could go red.

    In game 2 I have a two to one chance of being non-red against some unknown. Two to one are relatively good odds. I go red.

    But, according to Steve, “It’s fine to believe that the reds outnumber the greens, and fine to believe that the greens outnumber the reds. But when you’re able to believe both things at once, I start to worry about how you survive in the world.”

    Now I do appreciate his concern :-)

    He explains, “Most people choose to bet against red. Apparently they believe that the reds are outnumbered by the greens!”

    No, the reds are outnumbered by *all* balls. 2 to 1. The odds have changed. From 1 in 3 to 2 to 1. I do not believe both things at once, the odds have changed and I believe in the odds (and I do believe it’s entirely possible I misread the entire proposition :-)

    Back to COVID I was fine with the first coupla weeks of lock-down to give healthcare a chance to ramp up. Beyond that I believe in “laissez faire” … to some extent. I *will* keep a respectful distance from people whether I’m asked to or not. (People, especially the ladies, have always required this of me anyway, so, really, no change here.) And, if many to most are wearing masks I will simply conform out of courtesy.

    But, other than that, let’s put Darwinism to work. If I make the wrong mistake I won’t expect my neighbor to put a flower on my grave. She won’t be able to afford to after the economy collapses :-)

13. 13 13 Roger Schlafly
    July 1, 2020 at 5:34 pm

    I am not sure the lockdowns have saved any lives. The main stated purpose of them was to “flatten the curve”, so that hospitals are not overwhelmed. They may not affect the total number of infections over 2020-2022 at all. The flattening was apparently successful, because most hospitals are running at 10% capacity or less.

14. 14 14 joe henry
    July 2, 2020 at 6:29 am

    I’m with Roger #13, mandatory lock-downs may simply be prolonging the inevitable. X deaths over 6 months or X deaths over 24 months is still X deaths.

    Worse the lock-downs are responsible for some number of additional deaths due to depression, avoidance of healthcare for other issues, the downward spiral of seniors in facilities who have been cut off from family and socialization, job loss, exacerbation of poverty ad infinitum …

    To some extent even the radical social unrest across the nation might be attributable to the lock-downs.

15. 15 15 Harold
    July 2, 2020 at 7:13 am

    #12. The reds are outnumbered by *all* balls 2:1. But if you think there are more reds than greens (as revealed by your first bet), then the greens are outnumbered by *all* balls by more than 2:1. You have a 2:1 chance of winning if you bet red, but a greater than 2:1 chance of winning if you bet green.

    In the first game, I assumed you keep your stake unless the ball is either red or green. You may neither win nor lose if the ball is say purple.

    In the second game, you will certainly win or lose, as the ball either will or will not be red (green). The games are not the same, but if you want to win, then changing from red to green holds assumptions consistent, whereas sticking with red requires inconsistent assumptions about the number of green balls.

    The argument against this having any significance is that the assumption is just that, and we have very little reason to have any confidence in it. Changing the assumption, which we did not believe with any confidence, does not reveal very much about our consistency.

    #13. This point is reasonable. It is why I referred to immediate lives lost, and said we will not know for a while. It is likely that SARS-CoV-2 will become endemic. It probably will not disappear as did SARS-CoV-1

    If we get a vaccine, then there will be far fewer deaths overall. If we get a treatment there will be far fewer deaths overall. If we get neither, it is less certain.

    Given the stakes and the scale of the response, I think we will get a vaccine, although nothing is certain.

16. 16 16 joe henry
    July 2, 2020 at 12:13 pm

    Harold #15 thanks for making me think (Ouch!)

    It starts with Steve teasing those of us who were “silly” enough to bet *for* Red and then … *against* Red as inconsistent. But it’s not. It was never about color. It was always about the odds.

    The first game was about *bad* odds versus unknown odds. It’s a coin toss. Pick Red.

    The second game was about *favorable* odds versus unknown odds. Now I all but have to go *against* Red.

    Methinks Steve likes to play. It was never about color. I’m betting Steve’s followup will be BLM isn’t about color either :-)

17. 17 17 F.E. Guerra-Pujol
    July 3, 2020 at 3:41 pm

    Although I agree with Steve’s framework in his excellent essay–and about the unavoidability of trade-offs and the nonsense spewed by politicians–, even he has to admit that there is (and I quote) “so much uncertainty about the costs [of the lockdowns and mask policies] and the number of lives saved” that it’s impossible to tell whether the lockdown policies are worth the trade-offs. Given this uncertainty, deciding what the wisest course of action is or what public policy approach to adopt will depend on our priors, and our priors, in turn, will be shaped by our politics, right?

18. 18 18 William
    July 4, 2020 at 2:05 am

    Another measure of how much the average human life ought to be worth worth is to integrate GDP per capita as a function of year for all the years that an average person lives. I did a back of the envelope calculation and got a value much smaller than the $10,000,000 although I think getting that number would require factoring in future values of GDP per capita. If people valued there life less or more than the integrated GDP per capita, that would be sub-optimal. Of course this measure would depend on the economic growth rate, and so optimal risk taking should factor in said expected growth.

19. 19 19 Francisco
    July 4, 2020 at 11:50 am

    What would happen if the US government decided to pursue one policy and to pay each American $1,000 each if they didn’t catch the coronavirus Mr Landsburg?

20. 20 20 Harold
    July 4, 2020 at 6:52 pm

    #18 If we take a much more simplistic view and multiply current GDP per capita by life expectancy we get about $4M. That is in the same ball park as the VSL.

    This simplistic view “sort of” factors in changes in the value of a dollar over time.

    #16. I have had the benefit of spending much time pondering over Steve Landsburg’s previous rationality tests. They are not really his scenarios, but this is how I became familiar with this sort of problem.

    The Monty Hall problem is a fairly trivial example. The more complex ones are an examination of what one means by “rationality”

    Going back awhile, one was here
    http://www.thebigquestions.com/2011/01/05/another-rationality-test/

    It took me a good while to figure out what the probem was, and what the answer meant. Having wrestled with such problems as presented by Steve Landaburg, the current ball-in-a-jar problem seemed obvious, but I can reflect that it only seemed so because of a training in this sort of problem.

    It seems to me that we must accept that people are not always rational, because it requires training in these problems to arrive at the “rational” answer.

    That does not mean that people ae not rational per se, only that they are not rational in the von Neuman-Morgenstern system of axioms. I don’t myself see a problem with with the VnM system, so my conclusion is that people are not always rational, but that is not the only possible conclusion.

21. 21 21 arch1
    July 4, 2020 at 8:02 pm

    Thanks for the article, always enlightening. One nit: I think that store/restaurant/bar owners are only partly (not “fully”) incentivized to look after their customers’ interests, because it is hard to trace a mild outbreak to a given establishment given all of the many-to-many relationships in play.
    As a result, it seems likely that owners will tend to relax de facto distancing rules if this yields (say) 20% more customers coming in the door,even if this would result in a somewhat slightly higher infection rate inside.

    (I was going to pick another nit around possible collusion in the urn game, when I realized that a charitable readieng of “with the same urn” removes any benefits of collusion betwen the person who filled the urn and the person on the other side of my bet:-)

22. 22 22 Harold
    July 5, 2020 at 7:37 am

    #20. No collusion!

23. 23 23 James Roberts
    July 5, 2020 at 6:21 pm

    Thanks, Steve. I have distributed this to my colleagues.
    As to the text itself/organisation, you make a jarring shift when you start a new paragraph with: “Now let’s talk about COVID-19.”
    Market failure (externalities) are at the heart of the logic of government in a civilised society.
    When people can’t co-operate, there’s money on the table.

24. 24 24 JB
    July 6, 2020 at 9:51 am

    With regard to the efficacy of safety measures, I’ve long thought Joe Paterno and Mike Ditka had a great suggestion regarding concussions in football: get rid of the facemasks.

    https://www.nbcbayarea.com/news/sports/paterno-and-ditka-nfl-should-get-rid-of-facemasks/2088221/

25. 25 25 Harold
    July 7, 2020 at 8:05 am

    #23. Somewhat off topic, but I was looking at when US Football diverged from Rugby football. It is complicated, but apparently one big change is the difference between off-side and line of scrimmage, but the forwards pass is the obvious big difference. What I found interesting was that “formations” used to be allowed, and players effectively re-recreated early medieval warfare, with formations like boars-snouts and shield walls. It all proved too dangerous, so formations were banned.

    More on-topic, a study in the US comparing injuries in collegiate football and rugby found significantly higher injury rate in rugby.

    https://journals.sagepub.com/doi/abs/10.1177/0363546515622389?journalCode=ajsb#:~:text=Results%3A%20The%20overall%20injury%20rate,3.0%2D8.6)%20in%20rugby.

    I don’t know if this would be the same if we compared European rugby with American football. It is at least plausible that Americans grow up with American football and therefore play rugby in a more reckless manner that their European counterparts as they are used to having protection.