And the prize for the best online information system (category: supranational body; subcategory: open access) goes to...
The UN Food and Agricultural Organization World Agricultural Information Centre.
WAICENT has tonnes of information and data across countries and time, it is easy to find, and simple to use from what I can see. WAICENT is available at http://www.fao.org/waicent/
The UN and its affiliates are good for providing information. The core statistics department at http://unstats.un.org/unsd/default.htm is helpful for broad information on most macroeconomic-type things. The UN International Labor Organization http://laborsta.ilo.org/ is a source of information unlikely to be consolidated elsewhere but which is needed surprisingly often, like unemployment rates in developing countries.
The US Government websites used to be a good source of information for things like energy statistics, but I don't use them much nowadays. Students use the CIA World Factbook occasionally in their essays, so I hope that the authors don't distort the presentation and choice of data for propaganda reasons (more than is inevitable). It is a global public service, so thanks and please keep it that way, or outsource the writing to a less politically pressurised body. Like they are going to listen to this European!
The EU website and Eurostat are great for European information, but when I last looked didn't have a comparable global range.
Monday, 31 March 2008
Friday, 28 March 2008
Dull macroeconomics
The neo-classical synthesis of Keynesian and classical economics sounds jazzy. Then you look at the main arguments in it, and find that the really important ones are about how long shop owners take to adjust their prices and whether people guess correctly how quickly prices will increase next year. Really, sticky prices and rational expectations underpin modern macrotheory.
To check whether these are true, you can look at macroeconomic models consisting of several equations and then watch as these equations are substituted into each other, then simplified, then substituted and simplified again. A graph might be drawn with a couple of lines and a point indicating an intersection. The empirical testing will use linear algebra, which is like primary school adding up except in matrix notation which can make things just opaque enough to lead to constant head slapping about how an elementary mistake has been made which a seven year old could understand.
Black hole theory it isn't, and no-one would mistake it for genetic engineering. In much of mathematical science there is constant aesthetic excitement which dulls the sense of hard work or frustration, but economics feels like work much of the time, although the flip side is that the analysis is generally more direct.
Having said that, there are occasional papers which catch the eye, and writers who can be interesting whatever the topic. One of the first papers I read on my specialism growth theory (as a Masters' student) was entitled "Fable for growthmen", from a celebrated author whose presentation was entertaining, topic important, and apparently contained an important original result (though I don't recall noting it at the time, to be honest). Colourful presentations may not be suitable generally for journals - professional academics are anaesthetised to the topic and can fly through results so that the colour might be a distraction - but in some textbooks they can be worthwhile, and occasionally in journals can ease the entry for bored Masters' and PhD students.
To check whether these are true, you can look at macroeconomic models consisting of several equations and then watch as these equations are substituted into each other, then simplified, then substituted and simplified again. A graph might be drawn with a couple of lines and a point indicating an intersection. The empirical testing will use linear algebra, which is like primary school adding up except in matrix notation which can make things just opaque enough to lead to constant head slapping about how an elementary mistake has been made which a seven year old could understand.
Black hole theory it isn't, and no-one would mistake it for genetic engineering. In much of mathematical science there is constant aesthetic excitement which dulls the sense of hard work or frustration, but economics feels like work much of the time, although the flip side is that the analysis is generally more direct.
Having said that, there are occasional papers which catch the eye, and writers who can be interesting whatever the topic. One of the first papers I read on my specialism growth theory (as a Masters' student) was entitled "Fable for growthmen", from a celebrated author whose presentation was entertaining, topic important, and apparently contained an important original result (though I don't recall noting it at the time, to be honest). Colourful presentations may not be suitable generally for journals - professional academics are anaesthetised to the topic and can fly through results so that the colour might be a distraction - but in some textbooks they can be worthwhile, and occasionally in journals can ease the entry for bored Masters' and PhD students.
Disproportionately useful theories #5: central bank independence
In a competition for the theory with the most impact on economics and public awareness of economics over the last twenty years, the theory of central bank independence would be a strong contender. The idea is that governments should not set interest rates on a day-to-day basis, but rather tell someone else what inflation rate they want and then let that someone else set interest rates to obtain an inflation rate close to the target. Inflation can be controlled by interest rates (since with high interest rates, economic activity slows down because borrowing is expensive, and so does inflation), and the central bank (the government's bank) is the organisation to control inflation because they have access to such large amounts of money that they can influence interest rates unlike anyone else.
There are many design considerations for the central bank. How often should it report back to the government? What role should the government have? Is the arrangement democratic? Should the central bank only think about inflation or worry about the effect of its actions on output too? What sort of person should be head of a central bank?
The theory's importance was boosted by a number of factors - it followed a period of very high global inflation and a subsequent slump, it coincided and anticipated increasing concern with institutional arrangements in economic affairs, it is comparatively easy for policymakers to understand and implement the recommendations, and its subject is considered important in neo-liberal economics' worldview.
The modern theory received its present impetus in a paper from 1985 entitled "The optimal degree of commitment to an intermediate monetary target", which by virtue of its topic is arguably the theoretical paper with the most impact since its publication. It produces theoretical results in most of the major pure economic aspects of optimal design, it uses mathematical models without the reader feeling that the theory would collapse entirely if the assumptions are varied, and it has a high ratio of clearly derived results to discussion. If the paper's extreme influence was inflated by circumstance, it took full advantage of the opportunity.
There are many design considerations for the central bank. How often should it report back to the government? What role should the government have? Is the arrangement democratic? Should the central bank only think about inflation or worry about the effect of its actions on output too? What sort of person should be head of a central bank?
The theory's importance was boosted by a number of factors - it followed a period of very high global inflation and a subsequent slump, it coincided and anticipated increasing concern with institutional arrangements in economic affairs, it is comparatively easy for policymakers to understand and implement the recommendations, and its subject is considered important in neo-liberal economics' worldview.
The modern theory received its present impetus in a paper from 1985 entitled "The optimal degree of commitment to an intermediate monetary target", which by virtue of its topic is arguably the theoretical paper with the most impact since its publication. It produces theoretical results in most of the major pure economic aspects of optimal design, it uses mathematical models without the reader feeling that the theory would collapse entirely if the assumptions are varied, and it has a high ratio of clearly derived results to discussion. If the paper's extreme influence was inflated by circumstance, it took full advantage of the opportunity.
Wednesday, 26 March 2008
DR Congo's mining renegotiation
Reportedly (http://news.bbc.co.uk/1/hi/world/africa/7306325.stm), Democratic Republic of the Congo is cancelling or renegotiating many of its mining contracts, because "none of the contracts met international standards".
It could be a smart move. African mining deals can give foreign companies returns on their investment which are laughably high. The company only has to operate for a month and it has recovered in profits all of its original investment. OK, Africa's sometimes risky, but not that risky.
It could be a smart move... if the renegotiated deals are transparent. Otherwise, the new terms could be similar to the old ones, but with a different minister or civil servant pocketing a bribe, and investors thinking that the government is volatile and corrupt. The current Congolese administration has overseen a good macroeconomic performance given the extraordinarily difficult circumstances, and hopefully it will continue to improve the economy.
It could be a smart move. African mining deals can give foreign companies returns on their investment which are laughably high. The company only has to operate for a month and it has recovered in profits all of its original investment. OK, Africa's sometimes risky, but not that risky.
It could be a smart move... if the renegotiated deals are transparent. Otherwise, the new terms could be similar to the old ones, but with a different minister or civil servant pocketing a bribe, and investors thinking that the government is volatile and corrupt. The current Congolese administration has overseen a good macroeconomic performance given the extraordinarily difficult circumstances, and hopefully it will continue to improve the economy.
freerice.com
There's a test-your-vocabulary game at freerice.com which is free to play and if you get correct answers then the sponsors give 20 grains of rice to the world's poor. It has the World Food Programme's approval. You are graded and can move up to level 55 which is, like, language black belt. It claims people rarely get beyond level 48, but I bet some hermit has attained level 55 perfection despite only eating figs.
And in the next decade...
As I mentioned in my post last Friday, Botswana, Burundi, and Gambia seem to have outperformed relative to the predictions of classical growth theory, which estimates growth rates purely on the basis of investment, education, population growth, and starting income. Their growth must have been due to other factors - they seem to have had the economic "X factor".
On the other hand, they may have been, if not exactly lucky, then beneficiaries of circumstances or factors of a transitive nature. So I looked at the growth rates for all African countries over the period from 1991-2006 using IMF data. If Botswana, Burundi, and Gambia have some special features, then one would expect continued outperformance, making allowance for extreme events during the period.
In terms of absolute growth rates, Botswana was near the top of the performers in Africa, beaten only by two oil producers, two market economies, and a tiny island state. However, its growth rates were half what they were in the 1980s, so it probably was performing in line with its classical predictions. I've made the assumption that savings and education did not plummet between the two periods, but not checked, being busy and lazy and that.
Gambia was in the top 25 percent of performers in absolute terms in 1991-2006, and its growth rate picked up by one percent from the previous decade. Again assuming no huge changes (increases, this time) in savings and education, Gambia seems to have retained its growth outperformance.
Burundi was the third worst African performer in 1991-2006. But it was in conflict since 1993, so it is difficult to take lessons from its performance.
In summary: one piece of evidence saying continued outperformance; one piece saying a reduction to expectations; one piece which can't be used. I wouldn't like to build a theory on this evidence.
On the other hand, they may have been, if not exactly lucky, then beneficiaries of circumstances or factors of a transitive nature. So I looked at the growth rates for all African countries over the period from 1991-2006 using IMF data. If Botswana, Burundi, and Gambia have some special features, then one would expect continued outperformance, making allowance for extreme events during the period.
In terms of absolute growth rates, Botswana was near the top of the performers in Africa, beaten only by two oil producers, two market economies, and a tiny island state. However, its growth rates were half what they were in the 1980s, so it probably was performing in line with its classical predictions. I've made the assumption that savings and education did not plummet between the two periods, but not checked, being busy and lazy and that.
Gambia was in the top 25 percent of performers in absolute terms in 1991-2006, and its growth rate picked up by one percent from the previous decade. Again assuming no huge changes (increases, this time) in savings and education, Gambia seems to have retained its growth outperformance.
Burundi was the third worst African performer in 1991-2006. But it was in conflict since 1993, so it is difficult to take lessons from its performance.
In summary: one piece of evidence saying continued outperformance; one piece saying a reduction to expectations; one piece which can't be used. I wouldn't like to build a theory on this evidence.
Friday, 21 March 2008
Residual analysis
Residual analysis is the examination of the gaps between predicted and actual data after fitting a model. Suppose we said that inflation would be equal to the economic growth rate, and one year inflation is 5% and growth is 2%. Our inflation prediction was 2%, so we have a residual of 5%-2%=3%. The residual analysis is saying that the residual is large and the model is dire.
Classical testing of models examines the sum of the residuals (actually the sum of their squares, so they don't cancel each other out when added). The approach is a bit different in modern tests of more complicated fitting issues than just gaps between actual and predicted data. For example, successive residuals might all be positive or negative, which would mean that the model underestimates for many consecutive data points - such as when the data is large - and overestimates for another set of linked data points - such as when the data is small.
Many of the famous modern tests rely on feeding the residuals back into another model, and seeing whether that model is a good fit. If it is, then the original model has something suspicious about it. For example, with the positive-when-the-data-is-large residuals, then we would expect a model such as residual = data - 10 to be a good fit, since large data (above 10, say) would have a positive residual, and small data would have a negative residual.
Classical testing of models examines the sum of the residuals (actually the sum of their squares, so they don't cancel each other out when added). The approach is a bit different in modern tests of more complicated fitting issues than just gaps between actual and predicted data. For example, successive residuals might all be positive or negative, which would mean that the model underestimates for many consecutive data points - such as when the data is large - and overestimates for another set of linked data points - such as when the data is small.
Many of the famous modern tests rely on feeding the residuals back into another model, and seeing whether that model is a good fit. If it is, then the original model has something suspicious about it. For example, with the positive-when-the-data-is-large residuals, then we would expect a model such as residual = data - 10 to be a good fit, since large data (above 10, say) would have a positive residual, and small data would have a negative residual.
Subscribe to:
Posts (Atom)