The Importance of Being Earnest: How Superfreakonomics killed contrarianism

I missed out on the Crooked Timber book title contest (combine a classic title with a “How C did Y” subtitle in the modern manner) a while back, so here’s my entry. As regards earnestness, i’m riffing off Andrew Gelman, via Kieran Healy at CT, who observes “”pissing off conservatives” is boring and earnest?”

The main point, though, is that the fuss over the global cooling chapter in Levitt and Dubner’s new book is the first occasion, I think, where the refutation of specific errors has taken a back seat (partly because, in this case, it’s so easy) to an attack on contrarianism, as such. The general point is that contrarianism is a cheap way of allowing ideological hacks to think of themselves as fearless, independent thinkers, while never challenging (in fact reinforcing) the status quo. Here’s Krugman and Joe Romm, for example

I can certainly remember that I was once positively disposed to contrarianism. Trawling through the blog records, I can find

* A mixed review of Christopher Hitchens (on our side then), Letters to a Young Contrarian. If memory serves, I had a more favorable view of contrarianism, and Hitchens, before reading the book than after.

* A reference to “The worst kind of contrarian: That is, one who makes great play with contradictions in the conventional wisdom, does not put forward a coherent alternative, but nonetheless makes authoritative-sounding pronouncements on public policy.”

* A diagnosis of Richard Lindzen as someone who is “just an irresponsible contrarian as a matter of temperament. ”

To sum up my current view in a contra-contrarian style: “contrarianism” is mostly contrary to reality, the “conventional wisdom” is probably wiser than the typical unconventional alternative, and “politically incorrect” views are almost always incorrect in every way: literally, scientifically and morally.

90 thoughts on “The Importance of Being Earnest: How Superfreakonomics killed contrarianism

  1. @Alice
    No, Hoover put the tax rate from 25% (I said 24% earlier, which was incorrect) to 63%. Subsequently FDR set it at 79% and then 90%.

    Regarding your attitude towards taxation, I think this sums up how fair higher taxes on the rich are (and this is from Wikipedia, which, to my knowledge, is not part of the right-wing/libertarian blogosphere):

    “In 2007, the top 5% of income earners paid over half of the federal income tax revenue. The top 1% of income earners paid 25% of the total income tax revenue. Forty percent of Americans pay no federal income tax.”

    The least you could do is say “thank you” and stop demonising them so vociferously.

    The rich already pay more than their share – in fact it’s time for governments to drastically cut the spending and the bail-outs. Also, the proportion of GDP devoted to government spending in the USA is actually higher today than it was during the Keynesian heyday of the 50s and 60s. It is also substantially higher in Australia – in fact, Howard spent like a drunken sailor. You’re just mad that he didn’t spend on you!

  2. @Alice
    Don’t worry, I’m aware that Gerard didn’t say that – although I actually know some people who think that wars have that effect. I said that so that we all remember that it was supposed to be a time of great sacrifice, hardship and abstinence – and it also underlines my point about the high level of saving getting capital markets back into equilibrium. Much like South Korea during its development phase – lots of production, little consumption and lots of saving/capital accumulation.

    And that’s not justification for throwing me in the clink, since I didn’t make any trickle-down jokes… So there.

  3. My point, maybe not as obvious as I thought it would have been, was that it took the closest thing to a full-scale command economy to effectively end the depression.

  4. dd

    I am not doubting that economists have known about dynamic modelling and understood positive feedback (which is different from recursion) for quite some time. What puzzles me is that economists, as you say, keep trying to “predict” the unpredictable and keep using obviously flawed models. Economic systems as they currently operate cannot be predicted in detail anymore than the time we will see the next sun flare.

    Agent based modelling can give a probability that an event will occur which conventional models cannot. I look forward to the day when we have the commentators saying there is a 90% probability that the Australian dollar will reach parity with the USA within one month.

    Agent based models enable us to try out variations in the way agents interact and see the effect. In particular they make obvious the interactions that are causing instabilities. If we can change the interactions so that positive feedback does not occur then we will be able to predict using conventional modelling (which I find ironic).

    Even if you do not do the detailed modelling simply by thinking of the interactions and how they might change gives ideas on how to “fix” the instabilities and create economic systems that will have fewer instabilities.

    The instability in the money system is mainly caused by the rule that allows money creation (new loans) to be backed by money itself. This is a positive feedback mechanism and so we have all these other rules such as fractional reserves, capital adequacy, reserve bank interest changes etc to try to mitigate the effect of the positive feedback – but it would become easier to control if we could change the underlying mechanism or the agent interaction so that the positive feedback was unlikely to occur.

    In the case of money creation the system is biased towards the creation of money backed by existing assets. That is we normally only give loans (create bank money) if we can recover the loan money by seizing an existing asset. As money itself is treated as an existing asset this means the positive feedback of loans built on loans is likely to occur and in fact most loans – I think it is well over 90% are backed by other loans.

    However, what if we changed the agent interaction rule that made loan creation to build new assets cheaper than loan creation for existing assets? At the moment banks generally do not lend against future assets and so the builder of assets has to obtain money from savings and the mechanism is called equity. Equity returns have to be much higher than loan interest otherwise investors will not invest so we have this bias against investment in new assets.

    What if we changed this? What if we made it possible to get loans at lower interest if you promised to create a new asset?

    Agent based modelling would enable you to try this out. Conventional economic modelling would have difficulty because there would be other effects and we would have to change other interactions and other rules.

    For example we would change the way we control the money supply from using interest rates to the issuer of the currency deciding how much money to allow as new asset loans. Again the effect of all these new rule changes would be difficult to model using conventional methods.

    However, changing the rule for a particular asset class would stop asset bubbles in their tracks and we would make it less attractive for people to lend money backed by other money.

    To see one effect of such an approach come to my workshop this Saturday where I will outline how we can get to zero net emissions by 2020 with no increase in the price of energy simply by making loans for renewables or ways of saving energy cheaper than getting a loan backed by an existing asset such as a coal fired power station.

  5. Just received notification that the department of innovation… is possibly going to use non recourse, zero interest loans repayable from earnings to commercialise innovation instead of handing out grants.

    This is a very important innovation in policy that, if extended, to other government programs and if tuned to stop abuse, will result in beneficial changes to the way governments intervene in the economy. The government can issue as many non recourse loans as it sees fit without impacting its budget. It does not have to go out and borrow money to give grants and it does not have to raise taxes to pay for such programs.

    The impact of this approach on the ghg policy will be worked through at the Sat seminar mentioned at #27. I will also describe mechanisms to help prevent abuse of the system and outline how the system can be self regulating through market mechanisms.

  6. gerard,
    Are you seriously trying to maintain that, particularly in terms of consumer welfare, the depression ended in the US in 1941? Production in the US did return to pre-depression levels around then, but this was directed at blowing things up and serious quantities of destruction. It is (IMHO) to draw a very, very long bow to try to paint that as being, in some way, productive expenditure that advanced welfare.
    Individual welfare in the US did not return to pre-depression levels until the 1950s, when much of the wartime command-style economy had been abandoned.
    The Depression was not ended by the war – the war continued it.
    In fact there is plenty of evidence that the policies pursued by Roosevelt made the depression deeper and longer than it otherwise needed to be.
    A really good point is made by comparing the extent and depth of the Depression in Australia (which pursued then orthodox monetary and fiscal policy after it happened) to the US, which indulged in pump-priming and restrictive regulation. Perhaps you can help by checking which came out of the Depression (particularly in terms of consumer welfare) first.

  7. Sorry Andrew I have only just noticed your comment #15.
    Firstly, let’s deal with your claim that the economy troughed after 1933? As this goes against what the historical data is telling us, namely that US GDP bottomed in 1933, you are going to have to offer up some evidence to support this claim.
    Your comment that the trouble started with the founding of the Fed is a surprise, I wasn’t aware you were a policy crack-pot.
    Smoot-Hawley – Not financial regulation as far as I know. The March 1933 bank holiday is widely cited as a turning point in the crisis, you are going to have to offer up some evidence that not only was this not the case, but that it actually worsened the crisis.
    Before moving onto the New Deal, which was instigated after the economy had already troughed let’s deal with the topic in question. What were the specific financial regulations prior to 1933 which brought about the financial crisis? If you do insist on blaming the Fed, be clear as to the reasoning so we at least have something to discuss other than vague assertion

  8. Just to follow up your last comment as well Andrew, a major reason for the rapid recovery in Aussie growth in the 1930’s was the fact there was no financial collapse and much lower private debt levels.

  9. The common view among economic historians is that the Great Depression ended with the advent of World War II, not that it was prolonged by it. I assume the “great deal of evidence” you mention is as strong as your evidence that the CRA and Krugman caused the global financial crisis. You’re doing a good job of making JQ’s point about “contrarianism”.

  10. sdfc,
    Correct me if I am wrong, but Roosevelt became President in Janusry 1933, did he not? Therefore, even with the “height of the depression” being in 1933 (with subsequent double dip as Roosevelt’s policies bit) Hoover was in no position to tighten anything.
    On the contrary, though, the deficit under Hoover increased markedly.
    I would suggest that you look at Robert Higgs work (although I in no way endorse everything he says) on the recovery – it really happened after 1946 (not 1941) and only after many or most of Roosevelt’s measures had been rolled back.

  11. Andrew, this is becoming tedious however just for the exercise let’s go through things one last time. US real GDP shrank in each year from 1930 to 1933 before growing again in 1934. While we don’t have quarterly GDP numbers for the 1930’s we certainly have industrial production numbers and these suggest the economy began growing again sometime around mid-1933. As you acknowledge Roosevelt was not inaugurated until January 1933, I’ll leave it up to others to decide whether Roosevelt can be held accountable for the worst of the slump.

    As for Hoover tightening policy, once again this is a matter for history. I think you are becoming a little confused with the difference between cyclical and structural deficits. Let’s not forget nominal GDP fell about 50% between 1929 and 1933.

    The double-dip recession of 1937-38 came on the back of tighter fiscal and monetary policy. If you want to dish out some of the blame for the double dip recession on a premature fiscal tightening then I don’t actually have a problem with that.

  12. @sdfc
    sdfc – Roosevelt can no sooner be blamed for the slump than Rudd in Australia but the ALS make up their own history dont they????

  13. @jquiggin
    This looks an interesting book. However, most of the work it references is not tackling the problem the way I suggest. Most modellers think how agents interact then build their model.

    What I am suggesting and what I wish to be involved in (when I get the time and some spare change) is to start with the real world and model the interactions of a small group of economic agents. This modelling is done at the individual level. Of course most of their interactions will be with agents that are “virtual”.

    We would build the model so that it could be run until it was able to emulate actual activity over a period of time.

    That is, we would build the model from the real world and much like a taxonomists we would classify agents according to their characteristics. This classification will lead to species of economic agents which we can then use for our modelling.

    We would start with a few simple types. Individuals, groups of individuals, governments.

    The model would build up and would be continually tested against the actions of a few real economic agents. The model would be populated by mainly virtual agents.

    While we have a lot of economic transactions and there are a lot of individual economic agents the number does not matter once we have identified the types and how they behave given a set of input conditions and a set of characteristics. Economic agents while they might have a large number of transactions (say a bank) behave in the same way for all the transactions and we as individuals are the same. We are creatures of habit. This means that this approach to modelling is practical and it is my guess that we can get good results with a few simple models of economic behaviour and with few economic types.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s