Thursday, December 12, 2013

Statistics Norway Shows Creativity

I know, I've been picking on Statistics Norway (SSB) quite a lot. But isn't it my duty as a taxpayer in Norway? What I'm going to say this time is most probably too arrogant, taken into account that I'm an amateur, whereas SSB is full of professionals. I've never even seen a "macroeconomic model" in real life. But due to some combination of luck and skill, when it comes to house price forecasts during the last 12 months, I've beaten the KVARTS model SSB uses. That's where the arrogance probably comes from. Don't get me wrong: I have not published any accurate house price forecasts but, unlike KVARTS model, I have at least seen the possibility for price decline given the current fundamentals.

I will now simplify things (but so does KVARTS), and you should take this as my guess at what has happened here:

We got a "sneak peek" at KVARTS in June this year, when SSB invited an economist from a private bank to test the model. He wanted to see what, according to KVARTS, happens to house prices (hot topic in banks as well, it seems) if Brent oil price drops to $50 and stays there for the next three years. You might think house prices would drop? So did the economist, but this is not the case. Like him, you've most likely underestimated the robustness of Norwegian economy, as this would only reduce the house price growth to 2,1 % for the three-year period 2014-2016. Ok, this sounds ridiculous, you might think. Surely they have not taken into account the psychological effects? Wrong again. Without the effect of negative sentiment, prices would climb 6,3 %.

There you have it. That's KVARTS (version 06/13). Unable to see a house price decline even in a scenario many would label "doomsday".

Well, it so happens that the housing market cools down fairly dramatically and house prices start to decline during the summer, and this without any drop in oil price nor a visible drop in the absolute level of any other fundamentals. What do you do with a model that doesn't match reality? You scrap it, or you adjust it. Taken into account the "robustness" of KVARTS I described above, it probably doesn't take a small adjustment to get the house prices falling? Remember, the absolute levels of fundamentals have not really changed. But SSB doesn't have the time or the resources to overhaul the whole model. No, this calls for something like the opposite of deus ex machina ("diabolus ex machina"?). Enter the "strongly negative sentiment".

Here you have KVARTS, version 12/13. Finally able to see a house price decline, although only as a consequence of people's irrational fear and pessimism. It's very convenient, because it means that you don't really need to change your model for it to match reality, do you? The model is pretty much correct as it was, but the problem is with the reality; there are some irrational short-term fluctuations caused by strongly negative sentiment. In plain English it means "people have got it all wrong", and in plain Norwegian something like "folk har tatt feil". Because this sentiment doesn't match the real reality (?) suggested by the fundamentals, the effects will be short-lived, and soon enough we're back on the upward trajectory suggested by KVARTS.

SSB, please correct me if I got it all wrong?

What I think has changed in the economy, in reality, is the rate of change in the fundamentals. Like I showed in my previous post, SSB knows this - everyone knows it. It's called "utflating" in Norwegian and it's been going on throughout the year, not just in the autumn when the sentiment became increasingly negative.  It's the flattening out that increases pessimism among (heavily indebted) people. It's fundamentals driving sentiment (although this smells like "the chicken or the egg"). People are not stupid. When the growth slows and turns even a little bit negative, they get more pessimistic - for a reason. People got pessimistic because of a suddenly cooling housing market and real house price decline, and the price decline was due to other fundamental factors which even a real estate broker has admitted. They also got pessimistic when they heard (last winter?) that economic growth is slowing down in Norway.

To me it seems the change in the rate of change in fundamentals is clearly not captured even close to the extent it should be in KVARTS. The "bust" in boom and bust is totally missing here? Again I point out, as my amateur opinion, that a model heavily informed by what happened in 2008-2009 might be too robust, especially if that's the only real "bust" that falls within the data period. As far as I know, KVARTS uses data back to 1995, and so ignores the housing bust of late 80s and the recession that followed it.

Ådne Cappelen and Thorbjørn Eika at SSB, please read some Hyman Minsky, will you? Especially the parts where he talks about how a stable, robust economy invites instability by making people take oversized risks through oversized debt? It seems to me that in your model you have build the most robust economy ever, and if it doesn't call for too much leverage, and thus instability, then I can't see what would? I know, you can't incorporate these things in your model, even though you might think there's something to it. Well... Fuck the model? At least it should come with the same warning stickers cigarette packs have on them. "The Norwegian people, in aggregate, are probably smarter than KVARTS."

Feel free to share this, because I'd really like to hear some opinions from experts on all this. (Twitter: @catonyourface)

(I can't resist the temptation to float a "conspiracy theory" here: Was the time and resources needed to adjust KVARTS the real reason for SSB not giving out a forecast in September? The timing might be a bit too early, though.)


  1. Excellent post. I would really have liked to have a peak inside KVARTS, to see how it's modelled. Surely it couldn't just be some really simple regression analysis/fitting of the data from the past 20 years (aka the ruler method). Time to go back to school, SSB!

    1. Thanks, Kim! The only thing I can say without seeing the model is that it seems to be out of touch with reality. Does someone know if Norges Bank relies on the same model? Their forecast looked very similar

  2. For me, these models are totally useless from an epistemological viewpoint. You cannot falsify hypotheses or make future predictions based on historical data when it comes to human behaviour. The only purpose which these models can fulfil is to describe historical facts.

    Why? If one believes that human beings can learn, one irrevocably comes to the conclusion that causal explanations about human behaviour are not possible with the above described methods. The assumption which one has to make is the following: The nature of human behaviour does not change over time. Or more complicated, one assumes that a contingency of causes to effects exists and also a contingency of non-causes to non-effects. But this assumption is wrong if one believes human beings can learn.

    In the natural sciences, this assumption is more plausible because scientists face a different object area. You will (most likely) every time experience an oxyhydrogen reaction if you mix oxygen and hydrogen over a flame. But if you give a beggar on the street ten Norwegian Krones, he will maybe buy alcohol one time and bread another time. In the social sciences, you deal with conscious human beings who change their behaviour over the time.

    Consider this: If the so called "scientific method" is the best solution for dealing with these issues in the social sciences, why do economists not have the same success with it like natural scientists?

  3. Good points, Elias! Have you read Karl Popper? I haven't yet, but it's high on my reading list.

    1. Hei Huple's Cat,

      Elias is Eric; I just forgot to use my pseudonym yesterday. I read Popper a little bit, but not really into detail. If you want to read a good overview of the Popperian philosophy, take the book "Karl Popper" by Bryan Magee. Personally, I favour more the rationalist side and in particular Immanuel Kant and Ludwig von Mises.

    2. Thanks for the tip, E! I'm reading "John Maynard Keynes" by Hyman Minsky right now. (Have always around five unfinished books on my table, though). Can recommend this highly, learned already a lot and have just finished chapter 1!

    3. Sounds interesting. I currently read "The Ethics of Liberty" by Murray Rothbard. But the best book I read this year was not about economics. It was "The Art of Loving" by Erich Fromm. If you are interested in psychology, I highly recommend this one.

  4. What we should draw from this is that when house prices are very high it is all due to sound fundamentals and what is more those who already own property (the large majority) really do deserve all this apparent but as yet unrealised wealth that's fallen in their laps. It's natural, inevitable, and no it won't *ever* stop because Norway is special, any politician looking to be (re)elected will tell you this. Well done, you clever, clever electorate - you're rich and you deserve it.
    And when house prices fall? Well that's just negative sentiment - when prices fall it is, essentially, the people's own fault. It won't be long before some politician or other pops up to tell us that taking on a huge mortgage is a patriotic duty. The sales rate for new properties is really telling the story now.

  5. Why would you or any one else care to see how the model is made or by whom. The pseudo scientific model has a goal to show that in general all is fine just a small slow down but we will be back (probably) to the good all days. From my assumption the Norwegian economy as a whole has only one direction down. I know Norge best i verden but let's cut to the point. Norwegian strength and the economy is based on energy as well the pseudo oil fund.
    I am curious as hell what the outcome will the burst bring to the economy for the average (Johansen),having a 2 500 000 NOK house and a 300 000 NOK car and a cottage and most of it as debt. The action taken by the central bank with the benediction and help of politicians will apovrish the population. Have the pretty boys in suits haven't learn from the US mistake that started in 2006 and keeps running deep. No but it's Norway and we are different, really pull my hair spank me and call me stupid but I don't buy it.
    The Norwegian government had provided in 2008 (gave) 350 billion NOK (55 billion USD) in loans to banks - backed by mortagage bonds. In essence, it is a swap of mortgage-backed bonds for Norwegian Treasuries. That was in 2008 and now it's 2014 and I suspect that they will do similar thing.Just to put this into perspective, on a per capita basis, this would be 3500 billion USD if it were in the US. The amount is huge. Seems to me that the government/Norwegian taxpayers took on a significant liability in the housing market, to take that burden of bank's backs.-Simply put socialized looses privatized gains it's a excellent business, bank takes almost no risk.
    Energy question-Norwegian crude oil production peaked in 2001 and as of 2012 it has declined about 50% from the peak. This fact has been overshadowed by the oil price growth in recent years. This increase in prices has more than compensated for the decline in the amount of crude sold and delivered (and natural gas volumes), and has helped to maintain and grow annual Norwegian gross income.To run a balanced trade and cover domestic consumption Norway needs to produce 0.7 – 0.8 Mb/d at current prices. This level is presently expected to be reached within 15 years. Yes but this IS Norway and there is more oil in the sea waiting to be discovered...I hope so but for the information the Norwegian Shelf is the most surveyed on the planet earth and no MAJOR discoveries since the last 20 years. Also the pseudo gas reserves are not as big as told before very very exaggerated. "Norway’s largest newspaper “Aftenposten” covered on June 17th the entire front page with a headline stating “Rapid end to the gas”. The text began, “Snipp, snapp, snut. The Norwegian gas adventure will end much earlier than the authorities have stated. In ten years production will decline dramatically according to new calculations from Uppsala University.” (In Norway and Sweden one usually finishes children’s tales with the phrase “Snipp, snapp, snut, nu är sagan slut” - “Snipp, snapp, snut, now the story is ended)."
    Back to the real estate, when a family buys a home they will have a mortgage for 20-25 years right. So it's in the bank interest to keep the interest low and get as much of amortization:swap of mortgage-backed bonds for Norwegian Treasuries. So it's in their interest to keep on giving loans that are backed by the State( ex oil found) cause they are on the plus on the both end of the stick. PRIVATIZED GAINS SOCIALIZED LOSSES. SO keep on spending more debt it's good be confident, saving money is stupid.