I fired an email to Blanchard asking about any macro model that can generate data (in reverse, ie generating artificial data) which exhibits stylized facts and he pointed me out to the following paper he co-authored with Danny Quah:

“The Dynamic Effects of Aggregate Demand and Supply Disturbances”
http://www.master-ape.ens.fr/wdocument/master/cours/macro2/BQ%20AER89.pdf

If you read the paper above, you won’t find the sort of artificial generated data that exhibits stylized facts in the model described.

Stylized facts are given observational facts, ie, there are mathematical/statistical techniques used to reveal their presence/existence in the data. Revealing the existence of statistical properties in the data is a completely different scenario from a model that generates artificial data which reveals stylized facts after applying statistical magnifying glass (certain statistical methods ) on the data to see if it exhibits such behavior. Example, suppose that the stock price time-series exhibits periodicity (we know that it doesn’t but lets say), perhaps, a sinusoidal one (ie, cosine-wise or sine-wise periodicity). We can find this periodicity by applying FFT (fast-fourier-transform) to identify periodicity. The question to ask, if there can be some model (which can be formulated) that it can artificially generates data, whereby applying FFT to that artificial data reveals the presence of some periodicity in it. This is the question I was asking.

I sent a second message to Blanchard clarifying my question. I was asking for a model that generates data which exhibits stylized facts rather than a model or (statistical technique) that reveals stylized facts (as my example on FFT above). The 2 are not the same thing. So, I’m awaiting his response back. I’ve also emailed 2 other Macro-economists including Prof. Romer (from Berkeley), but I haven’t got a reply yet.

I have also sent an email to Damien Challet (who co-developed the minority games) asking the same question and here is his response.

Falafulu asked…
has this been done before (ie, a model reproducing stylized facts) in the field of neoclassical macroeconomics?

Damien Challet responded…
No, they can’t. Stylized facts are annoying for neoclassical models, to the point that they are considered anomalies in this field.

Damien also stated in his reply to my query that other agent-based models (in addition to MG) do reproduce stylized facts, such as from Lux, Bouchaud, Kirman, Iori, and others.

The reason I was asking in the first place, is because I read it in publications coming from complex system researchers and I just took their words for it, without any extensive knowledge in the macroeconomic literature. Surely they can’t be lying in their peer review papers because that is a no no in academic publications.

Until I’m being pointed out to macroeconomic model that can reproduce stylized facts, at this stage I’ll regard it as something that hasn’t been done in macroeconomics.

The DSGE and prior RBC literature was entirely premised on being able to generate economic stylized facts – the first chapter of any textbook talks about it. I can get Blanchard and Fisher out and start rattling off papers if you like – but I’d say just reading Chapter 1 indicates exactly where the focus is, and how important the ability to replicate stylized facts is to the discipline.

One of the major issues in terms of forecasting is the paucity of data – rather than the theoretical capabilities themselves. There is a timing issue (data takes time to get out and gets revised HEAVILY), and an unobserved variable issue, both of which hammer the ability to forecast in real time.

However, the ability to “generate economic/finance stylized facts” is hardly difficult – a model can be WRONG and still do that (I use the word wrong on purpose here – which is why its in caps 😉 ). This is why there is so much discussion on methodology, and how the theory fits into our conception of individual decision making – that is why expectations and choice will always be central factors, and they make up the main part of whatever macro theory will develop into.

Stylized facts in minority games with memory: a new challenge

The economics of Adam Smith had an undeniable moral dimension – modern economics tries the abstract from this dimension as far a practical and just describe trade-offs. So the comparison isn’t really appropriate.

@Falafulu Fisi

All interesting stuff. I think it mischaracterises macro a bit to be fair – but the key thing is that many of the concepts that are mentioned are things that were being discussed at point in time within economics. Remember, DSGE’s aren’t the entire discipline, they are one tool among many that economists use to try to understand trade-offs inherent in the economy.

I am surprised how much the paper states that economists ignore out of equilibrium dynamics when I remember having to sit through constant discussions on phase diagrams, multiple parteo ranked equilibrium, and the unsatisfactory nature of certain stability conditions while at school.

As economic theory develops (as it will, and has to – no-one ever claimed that the current version is the final version) the main thing that has to be kept in mind is expectations, and expectation formation. It is undeniable that we want to improve the description that is provided along microfoundations (an issue that is currently limited by computational issues more than by theory), but the expectations issue cannot be forgotten – and the papers appeal for adaptive systems shows a willingness to simplify the expectation issue that I think is risky.

The Economic Crisis is a Crisis for Economic Theory

“Invisible hand effect in an evolutionary minority game model”
http://arxiv.org/PS_cache/cond-mat/pdf/0403/0403499v2.pdf