A note on exchange rate regimes and macro outcomes

Tyler Cowen links to an interesting study by Andrew Rose, which found little difference in the macroeconomic outcomes of fixed and floating exchange rate regimes, during the recent crisis.  This surprised me a little, until I found out that the study excluded the eurozone.  Then it didn’t surprise me at all.  On the other hand almost everyone seems to have interpreted the results differently than I do, so perhaps I should be surprised.

It’s true that floating rate proponents like Paul Krugman and I have occasionally blamed the euro for the problems of Greece, Spain, Portugal, etc.  This might leads one to conclude that we predict that countries will do better, on average, with floating rate regime.  Indeed Krugman seems to have implicitly accepted that view.  But that is not the implication at all.  Rather the fundamental criticism of fixed rates regimes is not that they produce slower growth on average (impossible in natural rate models), but rather that they lead to greater variability.  The famous “one-size-fits-all” problem.

Here’s an easy way to see what I’m talking about.  All 50 American states operate with no monetary flexibility, as they are part of a single currency zone—the dollar. Ditto for the 18 members of the eurozone. Thus you would not expect the eurozone to perform differently than the US dollar zone, on any test of the effects of currency flexibility. Yet any study of fixed versus flexible exchange rate regimes would code the US as a “flexible rate” regime and each eurozone country would be coded as a “fixed rate” regime. At first glance, that would seem to imply that if flexible rates improve performance then the eurozone as a whole should do better than the average performance of its 18 members! Of course that’s impossible.

Instead, what you would expect is that the economic performance of floating rate regimes as a group would show less variability than the economic performance of individual countries within a fixed rate regime. Actually even that is perhaps an excessively rigorous test, because single currency zones tend to be composed of countries with similar structural characteristics. But even so I think it’s fair to say that this hypothesis holds up pretty well if you compare the eurozone with a group of floating rate developed countries including Switzerland, Sweden, Britain, Norway, Iceland, the US, Canada, Australia, New Zealand, S. Korea, Taiwan and Japan.  The eurozone may or may not have done worse on average than this group of countries (probably worse), but surely there’s been more variability in the eurozone. Countries like Germany have seen virtually no increase in unemployment, whereas unemployment in places like Greece, Spain and Portugal rose much higher than in any of the floating rate developed countries.

PS.  All fixed-rate currency blocs float against the rest of the world. There is no obvious reason why the monetary policy of fixed-rate currency blocs would be less effective than the monetary policy of individual floating-rate countries, at least when averaged over the entire currency bloc.

PPS.   When I disagree with the findings of studies, it’s not usually due to their formal technical analysis, but rather to interpretation.  Some of my recent “Mark Sadowski” posts are a good example. Another recent case like this occurred when Raj Chetty seemed to imply that the Oregon Medicaid study provided support for the Medicaid program:

Other economic studies have taken advantage of the constraints inherent in a particular policy to obtain scientific evidence. An excellent recent example concerned health insurance in Oregon. In 2008, the state of Oregon decided to expand its state health insurance program to cover additional low-income individuals, but it had funding to cover only a small fraction of the eligible families. In collaboration with economics researchers, the state designed a lottery procedure by which individuals who received the insurance could be compared with those who did not, creating in effect a first-rate randomized experiment.

The study found that getting insurance coverage increased the use of health care, reduced financial strain and improved well-being “” results that now provide invaluable guidance in understanding what we should expect from the Affordable Care Act.

Here’s one (anonymous) criticism of this interpretation:

With regards to Medicaid, Chetty also paints a surprisingly incomplete picture of the Oregon Medicaid experiment. As you will recall, Chetty is correct in pointing out that expanding Medicaid seems to have increased usage of health care, decreased financial strain, improved mental health, and improved self reported well being, but he, quite surprisingly given the caliber economist Chetty is, leaves out the less flattering (for supporters of the ACA) part of the study that found no statistically significant increase in objective measures of physical health for patients who received Medicaid.

At best, the Medicaid study was a mixed result for supporters of expanding the Medicaid program (which the ACA does quite dramatically). At worst, the study is a sad demonstration of how bad Medicaid (and perhaps insurance in general) is at improving objective physical health. Why Chetty presented this study as an unambiguous victory for the pro Medicaid crowd is a mystery to me (although I suspect support of ACA has something to do with it)?

In my view the Oregon Medicaid study provides support for replacing Medicaid with a new program called “Mediplacebo.”  I think the improvements in mental health identified in the Oregon study were real, and were important. But surely they can be produced at much lower cost. I know that every time I’ve had “cancer,” I’ve felt much better after going to the doctor and being told that I don’t have cancer. Under my plan, consumers would receive the same care provided to the uninsured for things like traffic accidents. For those health problems where the uninsured would not normally receive coverage, health consumers would receive a placebo.

Of course I’m joking, but there is a serious point here. Medicaid is a very expensive program. Unless and until it is possible to show clear physical benefits to health, the per capita spending on Medicaid should be scaled back to the levels that you see in other developed countries such as Singapore.  Ditto for Medicare.

We have a Cadillac health care system.  I don’t consider “no physical improvement” to be satisfactory, given the cost.


Tags:

 
 
 

38 Responses to “A note on exchange rate regimes and macro outcomes”

  1. Gravatar of Tyler Cowen Tyler Cowen
    22. October 2013 at 12:35

    While your points are very good ones, I am not sure they are the best way to read *the cited Rose study*, which focuses on the short run more than the long.

  2. Gravatar of TravisV TravisV
    22. October 2013 at 13:05

    Smart utilitarian: “We should aggressively reduce the growth rate of Medicare spending on unproven technologies.”

    Dumb utilitarian: “We should do everything we can to ensure that AS FEW PEOPLE QUALIFY FOR MEDICAID AS POSSIBLE.”

  3. Gravatar of rbl rbl
    22. October 2013 at 13:13

    I understand that the reason there are no “statistically significant increase in objective measures of physical health” is that that’s not what the research was looking for. There weren’t enough diabetics or high-colesterol participants to get a value indicating statistically significant results due to the treatment. The fact is, they received the treatment, and unless our current understanding of how to treat colesterol or diabetes is wrong, the expansion probobally improved their health. Basically, the study can’t show statistically significant evidence that insulin therapy helped the diabetics, but we already know that insulin therapy helps diabetics. Unless the medicaid recipients of insulin therapy have some inexplicable difference from other diabetics, the treatment probobally led to the expected outcomes.
    Source: http://angrybearblog.com/2013/05/medicaid-austin-frakt-aaron-carroll-and-kevin-drum-are-good-for-the-usa.html

  4. Gravatar of Matt McOsker Matt McOsker
    22. October 2013 at 14:22

    you must include balance of trade. a fixed-rate country can do just fine if they surplus trade surplus.

  5. Gravatar of Geoff Geoff
    22. October 2013 at 16:51

    The reason why the more decentralized floating exchange rate regimes are superior to the more centralized fixed rate regimes, is the same reason why the more decentralized private money production regimes are superior to the more centralized government money regimes.

    Decentralization minimizes variability, because it lessens the negative impact of every individual’s mistakes, which are of course inevitable.

    Finance is perhaps the most regulated industry on the planet, and it has incredible volatility.

    Electronics is perhaps the least regulated industry on the planet, and it has very little volatility.

    This is no coincidence.

  6. Gravatar of Geoff Geoff
    22. October 2013 at 17:05

    The country is on pace to accumulate more new debt under the 8 years of the Obama administration than what was accumulated under all of the other presidents in history…combined.

    The day after Congress extended the debt ceiling by a few months, the Treasury borrowed an additional, and eye-popping, $328 billion.

    All this debt acceleration is a direct consequence of the Fed’s monetary policy, since the Fed has to monetize an incredible amount of treasury debt in order to inflate the money supply by an amount needed to continue to prop up the already inflation-distorted economy. With that much free financing to the Treasury, the amount of Treasury borrowing concomitantly increases.

    Loose money is what is keeping Treasury prices from collapsing.

    The latest debt stand-off is only a precursor of what’s to come. Once interest rates on Treasuries rise sufficiently, the already set fiscal APOCALYPSE, which nobody can claim was unforeseen or unexpected, will become manifest, and the choice at the Fed will eventually be to either bail out the Treasury and hyperinflate, or watch the Treasury declare bankruptcy.

    For many years the “hard money gold bug” straw men have been warning of the consequences of what market monetarists believe is too little.

    Let nobody on this blog claim they didn’t know.

  7. Gravatar of Geoff Geoff
    22. October 2013 at 17:37

    TravisV:

    Consistent utilitarian: “If murdering ‘only’ 10 defenseless babies brings happiness to 100 million crazy people, we should do it!”

  8. Gravatar of Charlie Charlie
    22. October 2013 at 17:39

    The Oregon study showed that Medicaid improved most health care outcomes, but that many of the improvements were not statistcally significant. The tests, we know now, were underpowered. There just weren’t enough people with certain health conditions in the study to show an effect even though the point estimates were quite reasonable. It’s just too much nuance for Chetty to even try to convey and for no gain. “Anonymous’s” description is much more misleading. The worst kind of “cult of stastical significance.”

  9. Gravatar of ssumner ssumner
    22. October 2013 at 18:12

    Tyler, My argument applies equally to the short and long run.

    rbl, That may be the case, but until someone shows these physical effects it remains pure speculation. No one denies that some medical procedures improve health. And no one denies that some medical procedures worsen health. Hospital errors kill 100,000s of people every year. The net effect is probably positive. But that’s very different from saying the net effect of Medicaid is positive, as the alternative to Medicaid is not zero medical care. I’d guess Chetty hasn’t read Robin Hanson’s posts on medical care.

    Every time a new study confirms the Rand study it gets dismissed. But I’m still waiting for solid proof that health insurance helps physical health.

    Matt, Actually that’s not true, but it’s a common misconception. Look at the Netherlands. Or the US during the Great Depression.

    Charlie, I agree that there is way too much emphasis on statistical significance, but your comment is way off base. If there is no statistical significance you have NOTHING. It’s quite right that it doesn’t prove there was no effect (the study was too small), it doesn’t prove anything. But the bottom line is that the Oregon study found no statistically significant physical improvement in health. I think if you asked the average voter whether that finding was important, they’d give you a very different answer from Chetty. Feelings of “subjective well being” are all well and good, but they most certainly are not the reason voters support Medicaid.

    I’m surprised there is almost no reaction to my post on exchange rates. I’d consider it one of the half dozen most important posts I’ve done this year. Krugman would be foolish not to cite it, as it clearly supports his view of floating rates. And he was having trouble addressing Rose’s study.

  10. Gravatar of Mark A. Sadowski Mark A. Sadowski
    22. October 2013 at 21:11

    Prior to the recession large current account imbalances were common in the EU. All members with the exception of the Czech Republic, Denmark, France Italy and the United Kingdom had either surpluses or deficits exceeding 3% of GDP in 2007. In the majority of these cases these were deficits although large surpluses existed in the Low Countries, Austria, Finland, Germany and Sweden.

    Between 2007 and 2010 a dramatic adjustment in these balances occurred with nearly all of these imbalances moving towards balance. Only Luxembourg and Portugal had large imbalances grow larger over the period and the number of large imbalances (over 3% of GDP) declined from 22 to 11. A strong and statistically significant correlation exists between the positive change in the current account balance and the gap between actual GDP and trend GDP growth (10 year) that developed between 2007 and 2010.

    A single factor ANOVA test supports the conclusion that countries with currencies pegged to the euro suffered larger declines in growth relative to trend. Two sample t-tests assuming unequal variances suggest that countries with pegged currencies did significantly worse between 2007 and 2010 relative to trend than either countries in the euro zone or those outside the euro zone with flexible currencies. An ANOVA test and two sample t-tests show that current account reversals between 2007 and 2010 were significantly larger in countries pegged to the euro than in countries in the euro zone or those countries outside the euro zone with flexible currencies.

  11. Gravatar of 123 123
    22. October 2013 at 22:44

    I think the most important thing was said by Tyler Cowen:
    “One of the benefits of floating rates is that a country can crank up the rate of price inflation when it needs to, to offset negative demand shocks. If you are targeting an inflation rate, you are not doing this and that means you lose out on a potential benefit of floating rates.”
    I.e. during the period covered by study, inflation targeting has failed so badly compared to NGDPLT counterfactual, that macro volatility with IT was indistinguishable from macro volatily with fixed rates.
    We should still expect that floating countries should do better during crisis during the crisis caused by the malfunctioning of fixed rate systems, and this study does nothing to change this conclusion, indeed it proves that inflation targeting is better, as fixed rates work no better than floating ones when inflation targeting fails.

  12. Gravatar of Daniel Daniel
    23. October 2013 at 02:31

    Yo troll-face

    the choice at the Fed will eventually be to either bail out the Treasury and hyperinflate, or watch the Treasury declare bankruptcy

    When that happens (snicker), we’ll be here to watch you gloat how the magical forecasting power of Mises helped you make it through the apocalypse.

    Nobody’s holding his/her breath, though.

  13. Gravatar of Michael Michael
    23. October 2013 at 04:53

    Scott,

    The Oregon Study was woefully underpowered to show statistically significant effects on many the parameters they looked at. There are several reasons for this:

    1. It was not a true longitudinal study – the study looked at rates of various outcomes at two points in time (prior to enrollment and 1 year after enrollment) and compared whether the rates had changed. For example, proportion of patients with high blood pressure pre-experiment vs. post-experiment. However, they did not look at what happened with individual patients over time. This is sort of similar to your criticism of inequality data – that data looks at the income distribution at various points in time, but does not follow invididuals and see how their income changes.

    2. Outcomes like “average change in blood pressure” would have needed much larger populations to show an effect, because most people didn’t have high blood pressure. Any effect of treating high blood pressure would be observed only in those who had high blood pressure at baseline, but this was only ~20% of the the study population. This was even worse for diabetes-related outcomes because so few patients had diabetes at baseline.

    3. Only about 25% of those offered Medicaid actually joined.

    Clearly, this study did not prove the that Medicaid enrollment had any statistically significant benefits on physical health parameters. But absence of evidence is very different from evidence of absence.

    This study did not provide evidence that Medicaid enrollment does not improve physical health measures and cannot be considered credible evidence that Medicaid doesn’t work (as it has often been interpreted).

  14. Gravatar of ssumner ssumner
    23. October 2013 at 05:26

    Mark, I agree, but I very much doubt that causation runs from CA deficit to slow growth. Rather both factors reflect a third variable, economic structure. Those European countries with a certain economic structure happened to run CA deficits in the runup to the crisis. Note that Australia ran large CA deficits for decades, and did fine during the crisis.

    123, I completely agree.

    Michael, I agree with all that, but see my other replies. The burden of proof is on those who favor large expenditures on Medicaid. They need studies that show it works in terms of physical health. They don’t have them. Having said that, I favor a regime of universal health care. But how we design that regime depends very much on its physical health benefits. If there are none, a different design is appropriate. Something much cheaper, like Singapore.

    And Chetty’s comment was still highly misleading, even if everything you say is true.

  15. Gravatar of Michael Michael
    23. October 2013 at 06:01

    Scott, I will grant you that Chetty’s comment was misleading and that we should be after better evidence on Medicaid.

    BTW, the other Medicaid issue is that about half of Medicaid spending is not spent on the people we typically think of as “poor”. Rather, it goes to nursing home care for elderly who have spent down their assets(or shielded them, although that is getting harder and harder to do for most people).

  16. Gravatar of Charlie Charlie
    23. October 2013 at 06:51

    Scott,

    “If there is no statistical significance you have NOTHING.”

    I agree you have learned nothing with any confidence, but we absolutely did not learn anything about medicaid NOT improving health outcomes. The question is why you want Raj Chetty to report on everything a study doesn’t learn about. The Oregon Health study didn’t learn anything exchange rates, should he also report that?

    “But the bottom line is that the Oregon study found no statistically significant physical improvement in health. I think if you asked the average voter whether that finding was important, they’d give you a very different answer from Chetty.”

    Only if they were confused!! It’s almost as if you think the study found that the confidence interval around the health outcomes shows the effects were bound to be small. That is emphatically not what the study showed. In many cases, the point estimates were large and economically significant, but the standard errors were huge. That’s not the same as learning the effects were in some small range around zero. Bottom line, (get Bayesian) an informed reader would not have updated his views about the effectiveness of medicaid on health outcomes, there just wasn’t enough power.

    Please consider this: I’ve seen you analyze the stock market reactions to Fed announcements time and time again. Never once have I seen you note that the estimate was not statistically significant, even though you had one data point, your standard errors were infinite! I agree with that Scott Sumner, not this one.

  17. Gravatar of Matt McOsker Matt McOsker
    23. October 2013 at 07:07

    Scott – I never want things to be absolute, and balance of trade is but one factor. Though it should not be ignored. Italy and Spain have better export balances than Greece, thus they have not suffered as much as Greece. It should not be a variable that is set aside and ignored.

  18. Gravatar of Assorted links Assorted links
    23. October 2013 at 08:11

    […] 1. Scott Sumner on fixed vs. floating exchange rates. […]

  19. Gravatar of Mark A. Sadowski Mark A. Sadowski
    23. October 2013 at 09:49

    Scott,
    “Mark, I agree, but I very much doubt that causation runs from CA deficit to slow growth. Rather both factors reflect a third variable, economic structure. Those European countries with a certain economic structure happened to run CA deficits in the runup to the crisis. Note that Australia ran large CA deficits for decades, and did fine during the crisis.”

    But Australia didn’t have a pegged currency.

    There were five EU countries with exchange rate pegs in the run up to the crisis: Bulgaria, Denmark, Estonia, Latvia and Lithuania. The four largest CA deficits in the EU in 2007 were in Bulgaria, Estonia, Latvia and Lithuania (the BELLs). The four largest CA reversals from 2007 to 2010 were also in the BELLs. The three largest declines in RGDP relative to trend in the EU were in the Baltic States.

    The Andrew Rose study includes 83 countries that were “Hard Fixers” at least part of the time from 2006 to 2012, and yet he conludes here is little difference in the macroeconomic outcomes of fixed and floating exchange rate regimes, during the recent crisis. In the case of the EU this is clearly not the case.

    The countries with floating exchange rates were the Czech Republic, Hungary, Poland, Romania, Sweden and the UK. As previously mentioned Sweden had a large CA surplus in 2007 and continued to have one in 2010. Of the rest, Hungary, Poland and Romania had large CA deficits (3% of GDP or more) in 2007. By 2010 only Romania’s was still large. Nevertheless the CA reversal in these three countries was less than a third of the size of the BELLs. The decline in RGDP relative to trend in the six “floaters” was on average less than half as large than the five “fixers”, with Sweden having only a mild recession and Poland being the only country in the EU that did not have a recession at all.

    Now it seems to me that the huge CA reversals suffered by the BELLs was symptomatic of the huge decline in NGDP in those countries. Had they depreciated their currencies like the six “floaters” did relative to the euro between July 2008 and February 2009 (Poland depreciated the most, or by about 30%), this would not have happened.

    However, the question remains, why did the BELLs all have such large CA deficits before the crisis (and such fast growth in RGDP)? Was this completely unrelated to their peg with the euro?

  20. Gravatar of ssumner ssumner
    23. October 2013 at 11:25

    Michael, Good point about nursing homes.

    Charlie, Let me put it this way. If the Oregon study had found a statistically signficant impact on physical health, I GAURANTEE liberals would not have said “please ignore this result, the sample size is too small.” Well you
    can’t have it both ways. You can’t cherry puck the results you like and ignore those you don’t like. They tested for those things, so they should report the findings, even if inconclusive.

    Here’s another way of putting it. Suppose Chetty’s column had been in the NYT. I’ll bet if you asked NYT readers a day later what he had said, 99% would “remember” that he found Medicaid improved physical health, if they remembered anything at all.

    My reports on stock market reactions are far more significant that you assume. If you looked at daily market reactions the results would not be particularly significant. But if you look at the market reaction immediately after the data hits the market, the result becomes highly significant, as the average variability of asset prices approaches zero as the time frame approaches zero.

    I’m not saying everything I report is statistically significant, but when it’s not I almost always caution that the results are merely suggestive. I also look for evidence that the timing of the market reaction was linked to the timing of the policy announcement. Feel free to show me blog posts where I reported more significance than was warranted. I try to be careful, but probably err on occasion.

    Matt and Mark, Good points. I’m certainly not saying there is no relationship. My fear is that people might assume that this means CA deficits are “bad” and CA surpluses are “good,” which would be very unfortunate.

  21. Gravatar of Don Geddis Don Geddis
    23. October 2013 at 13:20

    Geoff: I’m thrilled, at you finally making a concrete, solid prediction: “the choice at the Fed will eventually be to either bail out the Treasury and hyperinflate, or watch the Treasury declare bankruptcy.” Welcome to the world of science!

    Now, my only question: when 5 (or 10? or 20?) years goes by, and we observe neither hyperinflation nor Treasury bankruptcy, will you finally admit that your economic theories are hogwash, and shut up?

    No, more likely you’ll find a clever way to rationalize your failed prediction. Hey! Why not give us a preview? I’m sure you can generate the rationalization now, for why it isn’t your fault (nor the fault of your theories), that your strong prediction failed to come to pass. What excuse do you think you’ll invent? Let us know!

    The one thing we know: no real world data, no matter what it is, can shake the faith of a true believer.

  22. Gravatar of Charlie Charlie
    23. October 2013 at 13:52

    Scott you said:

    “Let me put it this way. If the Oregon study had found a statistically signficant impact on physical health, I GAURANTEE liberals would not have said “please ignore this result, the sample size is too small. Well you can’t have it both ways.”

    But that is exactly wrong. You can absolutely have it both ways. That’s what low power means. If a test has low power, but still has statistical significance, it means even though the effect is measured with lots of error that the estimate is so large, we can still be confident it’s different that zero. The sample size is going to determine how wide the error bars are, but if with wide error bars the effect is still significantly different from zero, it’s completely reasonable to take that as strong evidence the effect is not zero.

    Yet, the other way doesn’t work. If the effect is economically meaningful, but the confidence interval is very large, then you can’t conclude much. Possibly the effect is very economically meaningful, possibly its zero or has the opposite sign. We just don’t know.

    “But if you look at the market reaction immediately after the data hits the market, the result becomes highly significant, as the average variability of asset prices approaches zero as the time frame approaches zero.”

    First, I don’t think I’ve ever seen you try to put a confidence interval on your posts. Have you ever?

    Next, I’m willing to accept the implicit assumption you’ve made that the event being studied doesn’t change the variance (which allows you to estimate standard errors to begin with–otherwise you have infinite errors and only 1 observation).

    Last, it’s true that variance of returns go to zero as the observation period gets smaller, but stock returns at high frequencies are highly non-gaussian. They have high skew and fat tails. Traditional t-tests are misspecified. These problems are surmountable, but I’ve never seen you try to surmount them.

    Again, I’m not saying you are wrong. I’m saying you have a lot of valuable things to say even though you don’t compute confidence intervals.

    “I’m not saying everything I report is statistically significant, but when it’s not I almost always caution that the results are merely suggestive.”

    So Raj is actually being much more cautious than you. In a Scott Sumner world, he could have said, “additionally, the evidence suggests that Medicaid helped lower the percent of patients with high blood pressure and high cholesterol, as well as improve markers linked to the health of patients with diabetes, though this improvements did not reach the level of statistical significance.”

    I think not saying anything is much more cautious and less misleading, but maybe this is somewhere we just have disagreement. To me, if you have a wide confidence interval that includes big effects, small effects, zero effects and negative effects, just don’t report it to lay people as it will probably confuse them. But you do indeed report such suggestive evidence all the time to your readers, and I appreciate the discussion.

    I wonder how many people who read your post or Anonymous’s post would, when asked the next day, answer that the effects on physical health were “small” or “near zero,” rather than “possibly economically important but measure with a great deal of error.”

  23. Gravatar of Felipe Felipe
    24. October 2013 at 04:55

    Scott:

    The burden of proof is on those who favor large expenditures on Medicaid. They need studies that show it works in terms of physical health.

    Hmm, “improving health outcomes” is not the only objective of Medicaid; subsidizing (a certain group of people’s) health care is of itself an objective too.

    Not affording health care at all is only the most extreme form of “health care is too expensive”. Less extreme versions include picking into your savings, selling assets, stop doing leisurely stuff, taking an extra work shift or borrowing money from friends/family. These less extreme forms would presumably not show any physical health improvement, yet I would venture avoiding those versions are also objectives of Medicaid expansion proponents.

  24. Gravatar of ssumner ssumner
    24. October 2013 at 07:32

    Charlie, You said;

    “But that is exactly wrong. You can absolutely have it both ways.”

    If you’d think about what you said here, you realize it can’t be right. Otherwise you could go into study X with Z prior belief. Study X could shift Z in one direction but not the other. That’s clearly impossible. No study (going in) has the potential to make it more probable that Medicaid is effective, but does not have even the possibility of making it less probable. That violates the laws of statistics.

    Filipe, I completely agree, but that’s not how Medicaid is being sold to the public.

  25. Gravatar of Charlie Charlie
    24. October 2013 at 09:02

    “If you’d think about what you said here, you realize it can’t be right. Otherwise you could go into study X with Z prior belief. Study X could shift Z in one direction but not the other. That’s clearly impossible.”

    You are all over the place now. First you want to read a hypothesis test, “If there is no statistical significance you have NOTHING.” Now you want to get Bayesian, which is exactly where I started. An informed Bayesian interpretation of how Medicaid effects physical health outcomes wouldn’t have moved the priors very much, because the standard errors of this study were large. But the little bit they did move, would NOT have moved the estimated effect of Medicaid on physical health towards zero, because of the RESULTS of the study. The point estimates showed positive effects on physical health around the size expected by informed priors.

    If you came in to the study with uninformed priors, the point estimate (your best guess) would still be that medicaid improves health outcomes for patients with diabetes, high blood pressure, and high cholesterol. It’s just that you would have a lot of uncertainty about those guesses (the distribution around you point estimate is wide–and includes zero).

    You are imagining some other study with different results that should move us towards the belief that medicaid has zero effect on health outcomes, but that study doesn’t exist.

    Consider the example you gave, statistical significance doesn’t compare how close Z is to X, it compares how close Z is to zero. In this study, Z was approximately equal to X, but Z and X, together, were close enough to zero that given the low power of the study, we couldn’t reject that Z is equal to 0.

  26. Gravatar of Charlie Charlie
    24. October 2013 at 09:59

    One more point said a slightly different way using hypothesis testing. The hypothesis test (to use your example) is “Is Z different from zero?” The answer is there is not statistically significant evidence that it is. We could also ask, “Is Z different than X?” The answer is “there is not statistically significant evidence that it is.” Yet, you want people to revise X toward zero, why?

  27. Gravatar of Geoff Geoff
    24. October 2013 at 17:37

    Don Geddis:

    You arbitrarily define “science” to be restricted to history.

    Mathematics and formal logic are not empirical, but they are fields of inquiry into the real world.

    “Now, my only question: when 5 (or 10? or 20?) years goes by, and we observe neither hyperinflation nor Treasury bankruptcy, will you finally admit that your economic theories are hogwash, and shut up?”

    No, for the same reason that I will not believe that just because a person chooses to go into a libtrary for 364 days in a row, that he will allegedly be forced by some law of nature to go into the library on the 365th day.

    My economic theories are not predictions. They are logical constraints.

    When I said that the Fed must choose to either hyperinflate, or let the Treasury go bankrupt, is my bet on the future, with no time restriction. Economic history depends on human choices, not inevitable laws of nature with a set time horizon.

    “The one thing we know: no real world data, no matter what it is, can shake the faith of a true believer.”

    You’re right. No matter what you observe, I will never shake my conviction in the Pythagorean theorem. If you ever believe you have observed a plane right triangle the relations of which violate the theorem, I will trust the theory more than I will trust your claims.

    The same thing is the case with economics. I am only ever certain on the logical constraints of economics. I do not have 100% certainty in predictions of the future that are empirical.

  28. Gravatar of ssumner ssumner
    25. October 2013 at 05:39

    Charlie, Alright, if the parameter estimates were as expected then you are right. Can you produce a source stating the parameter estimates were as expected for overall physical health?

  29. Gravatar of Daniel Daniel
    25. October 2013 at 07:26

    Shorter Geoff

    My predictions may never come true, but they’re still correct

    What a moron.

  30. Gravatar of Charlie Charlie
    25. October 2013 at 08:23

    Scott,

    I’m not sure there is something defined and measurable called “overall physical health.” I think you’ll have to look measure by measure. The study does compare its point estimates to what would be expected and discuss power. The passage below discusses the expected effect from more diabetes patients taking medicine, as well as the effect of medication for hypertension.

    “Nonetheless, our power to detect changes in health was limited by the relatively small numbers of patients with these conditions… The 95% confidence intervals for many of the estimates of effects on individual physical health measures were wide enough to include changes that would be considered clinically significant “” such as a 7.16-percentage-point reduction in the prevalence of hypertension. Moreover, although we did not find a significant change in glycated hemoglobin levels, the point estimate of the decrease we observed is consistent with that which would be expected on the basis of our estimated increase in the use of medication for diabetes. The clinical-trial literature indicates that the use of oral medication for diabetes reduces the glycated hemoglobin level by an average of 1 percentage point within as short a time as 6 months.15 This estimate from the clinical literature suggests that the 5.4-percentage-point increase in the use of medication for diabetes in our cohort would decrease the average glycated hemoglobin level in the study population by 0.05 percentage points, which is well within our 95% confidence interval.”

    http://www.nejm.org/doi/full/10.1056/NEJMsa1212321#t=articleResults

  31. Gravatar of Charlie Charlie
    25. October 2013 at 08:28

    I learned about these issues from Austin Frakt on econ talk:
    http://www.econtalk.org/archives/2013/05/frakt_on_medica.html

    Here is a Austin Frakt blog post:
    http://theincidentaleconomist.com/wordpress/updated-power-calculation/

  32. Gravatar of ssumner ssumner
    26. October 2013 at 06:55

    Charlie, That quote doesn’t back up your claim. There’s no data on other indicators such as life expectancy. Just cherry picking a couple indictors where the sign was “correct” doesn’t show what I thought you were claiming. Not saying you are wrong, but I’d need to see much more evidence.

  33. Gravatar of Charlie Charlie
    26. October 2013 at 08:49

    “There’s no data on other indicators such as life expectancy. Just cherry picking a couple indictors where the sign was “correct” doesn’t show what I thought you were claiming.”

    Scott these are the measures of physical health that Anonymous was talking about that showed “no significant effect.” I’m showing that the point estimates were close to what we’ve observed in clinical trials. And I didn’t cherry pick a couple of indicators. You asked for a source comparing observed effects to expected effects. This is what is reported in the study.

    I’m not sure you understand the study. This is two years later. How could you possibly look at life expectancy? You obviously need a long-term study for that. The authors don’t even look at mortality, people just aren’t dying in that amount of time.

    In that last two posts, you’ve asked for measures of “overall physical health” (which isn’t defined and measurable thing) and proof that a two year study changed long-term life expectancy. I’m not sure that you know what we’re arguing about. Perhaps it’s time you read the original study, something I’m sure Raj Chetty has done.

  34. Gravatar of Charlie Charlie
    26. October 2013 at 09:22

    ^I was rude above. I don’t respond well to accusations of cherry-picking, and I found your comment quite frustrating. I apologize.

    Let me just say, there is no overall physical health measure used in the study. The media created “medicaid has no effect on overall physical health” as a short hand for several individual health indicators did not show statistically significant improvement.

    I don’t think I ever said “overall physical health,” but if I did I just meant it as short hand for most of the individual point estimates showing improvements (of course with wide confidence intervals).

    We care about these measures, because in other studies they have been linked to things like mortality and life-expectancy, but linking medicaid directly to mortality or life-expectancy would not be possible–thinking about why brings us back to the concept of statistical power.

  35. Gravatar of Geoff Geoff
    26. October 2013 at 10:30

    Daniel:

    Aww, you sound mad.

    Actually, my prediction of a sluggish economy post-2008 did indeed come true. I predicted that because the Fed decided to intervene to stop needed corrections, the economy would remain sluggish for many years.

    But this is neither here nor there. You’re just ignorant about the science of economics as it is distinguished from guesswork of the future.

    The law of marginal utility is not an empirical prediction. It will always be true, by virtue of the nature of action. Any action taken whatsoever, and hence any possible future, will always be constrained according to the law of marginal utility.

    What you are doing is using a whole slew of a priori assumptions, falsely believing they’re empirically derived, and then you believe that because history is consistent with those assumptions, they are somehow the only valid assumptions. But that’s astrology. You’re a bookeeping astrologist, and nothing even close to an economist.

    Economics proper is a system of formal logic as it applies to action. Austrian economics for example does not make any predictions whatsoever. It can only be shown as wrong through ratiocination. It cannot be shown as wrong empirically, because it is an empirical science.

    I guess I can assume that you will again avoid actually engaging the arguments presented to you. You’re scared, I can tell. Your fear shows itself as hatred and anger at me.

  36. Gravatar of Charlie Charlie
    26. October 2013 at 11:43

    In response to an update by Anon Ymous, I wrote two posts to try to carefully explain my position. It seemed in principle we were coming to some agreement about how a study should be read.

    http://badoutcomes.blogspot.com/2013/10/what-is-statistical-power.html

    http://badoutcomes.blogspot.com/2013/10/a-bayesian-analysis-of-oregon-health.html

  37. Gravatar of Anon Anon
    26. October 2013 at 14:39

    Scott,

    Thank you for linking to my post, as it has greatly improved my site traffic. It was also rewarding to see you link to my blog as I am a regular reader of this blog and a big fan. You changed my views on monetary policy quite dramatically (I switched to supporting NGDP targeting about a year ago).

    I respond to Charlie in these two posts, where the discussion of the Oregon study continues:

    http://theanonymouscommentator.blogspot.com/2013/10/more-on-oregon-medicaid-study.html

    http://theanonymouscommentator.blogspot.com/2013/10/even-more-on-oregon-medicaid-study.html

  38. Gravatar of ssumner ssumner
    27. October 2013 at 07:17

    Charlie, If Anon’s facts are correct then your criticism of my “cherry-picking” charge was off base. He showed some examples with different results.

    Obviously you two know more about this than I do so I’ll follow your debate with interest.

    One thing that makes me suspicious is that I’ve met people who claim the famous Rand study from many decades back showed health insurance improved health, when it did not. So I’m very skeptical of health insurance from a health perspective (not from a financial risk reduction perspective.)

Leave a Reply