BONUS BALTICAPP - Race Against Eutrophication Blog
29.11.2017 15:05Theoretical vs. empirical/applied approach
It has become apparent for me recently that there is a great dividing between theoretical and empirical/applied approaches in science. Particularly in economics this dividing seems to be very strict. This was made clear to me when I started an Advanced Econometrics course in PhD program for Economics this fall. We were told that this course is purely theoretical and if there are anyone interested more in applied side of things they may consider hard if the course is too difficult for them. We were told that particularly PhD students in so called applied sciences should probably dismiss. I decided to not to dismiss and it was a good call since so far studies have been going well, although I’m definitely an applied researcher and as such not a “real” researcher, as it was made clear to us.
Ever since I’ve been thinking about this division and classification of the researchers. In my opinion theoretical and applied sides are both rather important. It is clear that there has to be solid theory behind an applied work. But it appears equally clear that good theory is worth of nothing if there is no empirical evidence to support the theory or if there is no application for the theory. In addition, new ideas for theory development may originate form surprising empirical findings. Also, applied work may show that the theory does not hold in practice. Therefore, it seems that there is justification for both sides and those can actually benefit from each other.
Thus, it seems rather counterintuitive that theoretical researchers don’t give much appreciation for applied researchers. In econometrics and statistics this might have something to do with the fact that the methods are often applied in a bad manner and important theoretical aspects are ignored. If theoretical researcher reads such empirical paper that is published in a journal, it is understandable that the theoretical researcher becomes upset and feels that his/hers work is not understood.
In many cases and in many papers statistics are done wrong. The problem might be that statistical and econometrics courses are typically either strictly theoretical or strictly applied. Then we learn advanced theoretical things in one course and some examples and some coding on another course. The problem is that the connection between theory and the empirical application might be left incomplete, to say at least. For example, in our Advanced Econometrics course we struggle mostly with the proofs of the theorems. We have learned how to derive distributions of the test statistics and to show the asymptotic consistency of the parameter. But how to apply these learnings with the real data? In the other hand, I took another course that was purely applied. There we were given some data and some codes and examples, but the theory behind the examples was not discussed at all. A good course would really combine these two worlds, because clearly those are not separate.
I don’t how is the situation in other fields, but at least in Economics I see an opportunity for improvement in how theoretical and applied sides of science should be thought. This would narrow the gap between the field work and work that is done in theory labs. It would probably decrease the hatred between the deputies of the sides. Also the books should entertain the theoretical justifications and empirical applications to really show the students/researchers how the theory is applied correctly to real data.Best regards, Matti Sihvonen
20.07.2017 12:00EAERE 2017
In this blog I will tell about the EAERE conference held in Athens. First I must emphasize that it was really hot out there. I have never experienced such a hot temperatures (35-45 degrees) in my life before. Luckily the air conditioning worked just well in the conference places and all the other facilities were also very well handled. Overall the conference was a good and learning experience. I witnessed some great presentations and speeches. I guess that the best presentations (among those I saw) focused on micro-econometric data analysis and intertemporal optimization. I guess that it was particularly beneficial to see the overall level and current trends in environmental economics. It appears that macro models and everything related to climate change is now cool. Also behavioral and network economics appear to be very hot fields right now. My subject of interest, namely empirical agricultural production analysis and agri-environmental modelling seems rather uninteresting subject currently, to say at least. But that’s ok, I keep calm and continue working with my PhD anyway. At least I am still excited about it (now that I’m working with the second paper)!
There was also one very good plenary speech, which handled multiple layers of uncertainty, including model selection uncertainty. I found this particularly interesting, of course, because I have been struggling with uncertainty (regarding particularly model selection) issues for so long time now. I actually started to loose my interest in structural uncertainty entirely. But now it seems that it might actually be the key to get my first paper published. I got this idea that perhaps I should reframe the problem so that the focus is, again, shifted to these structural uncertainty issues within the optimization problem. It would be absolutely perfect if I could find a way to formulate the profit maximization problem analytically so that the parametric and model selection uncertainty would be taken into account explicitly. This, however, might be very challenging and perhaps beyond my abilities, given the time limits at least. But in any case, I should stress the structural uncertainty, because it appears to be generally interesting topic after all; for a moment I thought I was the last person on earth interested in it.
My presentation in the conference went just fine. Needless to say I was very nervous about it. The audience wasn’t large, which was just fine. One reason for the lack of audience, among the fact that nobody is generally interest in empirically bio-economic modelling in the context of agriculture (or fertilizers (!)), might have been that it was a last parallel session before the dinner and the busses started to leave just after the session. I guess most of the people went to hotel to rest and get ready for the evening. Anyway, I got some comments for some Chinese senior researcher, which was very nice, although I couldn’t understand her comments at the moment, because of the very challenging accent. However, after reprocessing the echoes in my ears for a while I understood what she said. Anyway, the comments were very trivial and I’m a little bit upset that I couldn’t respond to those at the moment, but better luck next time perhaps. I also was a discussant for a one presentation in our session. I think it went well and the situation resolved rather nicely in the end. The dinner, by the way, was great and we had very good time. Even the head master of the Greek bank was there to give his speech, which was super funny (the speech I mean).
I must say that I didn’t seize the opportunity to network with foreign people at all, which was a little bit unfortunate. The thing is that there was such a huge amount of Finnish environmental economists there that most of the time we just chatted with each other in some isolated corner. And during the dinner we also had Finnish tables. Luckily some German researches shared a table with us in the end. Thus, some international communications did take place, but not much. Anyway, it was a very nice event and all. I really hope that I could participate the next one also (which will be held in Gothenburg), but it will be much more unlikely because it will be a world conference so the competition will be much tougher. I know for a fact that the second paper will be much more interesting in terms of environmental economics than the first one, but will it be enough, remains to be seen. It would be nice to see the Gothenburg because it is capital if Scandinavian Heavy Metal!
Thus, Rock on and see you next month!
Bets regards: Matti Sihvonen
05.06.2017 09:26Upsides and downsides, mostly downsides..
We had a very interesting PhD course on econometrics in Bergen Norway. We were introduced to modern methods of empirical data analysis including instrumental variables, differences-in-differences and regression discontinuity. The topics of the examples were from environmental and resource economics and included themes such as relationships between conflicts and resource discoveries, price shocks and civil conflicts, the political resource curses and environmental speed limits in Oslo. The course took three days. During the visit in Bergen we also got an opportunity to explore the beautiful city (or village). The surrounding mountains were particularly impressive. It also became apparent that the Bergen is a rather rainy place; there is approx. 300 rainy days a year, we were told. It also became clear that Norway is a very expensive country. The oil money seems to increase the prices beyond the level that at least we Finnish people would consider reasonable. The get the points from the course we have to hand in a home assignment by the end of the September. In this assignment we are asked to get some data related to our PhD research and to explore the data via some method introduced in the course. Then we are asked to write a paper about the data, method applied and the results. Thus, if data are good, one could expand this home assignment to a real paper. This is of course what I am planning to do if I get good data. Therefore the quest for data is now one of the priorities during the summer. I was thinking about data regarding some policy or regional shift. Thus, for example, data could be about an effect of a subsidy policy on different agricultural regions. Then I could apply regression discontinuity method for the data analysis.
We also had another PhD course, namely a project management course. However, although the lecturer was very good and inspirational person, I did not found the course very useful. I took it mostly as a language improving exercise than anything else since I have no interests in project management or managing in general. Of course it is useful to be able to manage ones own work but I don’t think that getting a course on that is very useful. Anyway, I’m sure that some participants did found the course useful and we did have some delightful conversations during the course.
There was also a third PhD course, namely Economic Growth and Natural Resources. We had to hand in some home assignments and then there was an exam. This was very interesting course because the topic was in the field of macroeconomics, which I am not familiar with. Therefore it was good introduction to the unexamined territory. Macroeconomics is good for environmental economist because most of the macroeconomic analysis is dynamic. Therefore I might take another macro course in September, which focuses on dynamic optimisation methods. I think it might be particularly useful for me.
It has also become clear that to publish a paper is rather hard task. We got a refusal from a first journal because “the article did not fit the scope of the journal”. The refusal, however, might have been a fortunate thing because the journal, namely the Environmental modelling and software, was probably not a good journal for the article because in the first paper we did not model any environmental aspects. The first paper discussed simply agricultural system consisting of nitrogen and phosphorus yield responses and the development of soil phosphorus. Therefore the scope of the article is clearly agriculture and not environment, although those two scopes are quite related. Only after the refusal I started to understand that the scope of the journal is rather important thing. I started to examine the potential journals for our paper. During this examination I realised that our paper is kind of a weird mixture of data analysis and optimisation. Therefore it might be difficult to find a journal that would consider the paper a good fit for the scope. Anyway, I did submit the paper for another journal and I am now waiting for a refusal or some other answer. I just hope that it will be accepted to some journal since the second paper is a direct extension for the first paper.
The refusal also reminded me of the saying of Hannu Vartiainen after the first Microeconomic lecture: “the PhD work consists of upsides and downsides, mostly downsides”. It certainly is so that the work has been mostly downsides and there have been times when I have considered calling my model a setback-model because there have been so many setbacks down the road.
The second, and actually the most important priority during this summer, is to get a second paper ready. I have been working on it now rather intensively and I think it is coming together quite nicely. So far the analysis has been theoretical, which is a nice change from a mainly empirical and numerical analysis carried out in previous paper. However, the next step is to include leaching functions for nitrogen and phosphorus into the numerical analysis. Therefore the scope is wider in this paper since it will discuss agricultural and environmental aspects via bio-economic modelling. As such the second paper will most likely be easier to sell to journal (compared to the first paper). But, it remains to be seen.
Meanwhile, we still wait for the summer to start here in Finland.
See you next time,
Best regards: Matti Sihvonen
14.04.2017 11:50The lost art of writing (and speaking)
Things have been moving quite nicely in recent times. First of all, we got the first article ready and it’s now submitted to a journal. I must say that it was rather challenge for me. The actual writing process proved to be very difficult in the end. I had some real difficulties to put a discussion, conclusions and abstract sections together. This may be due to the fact that we, students in economics, can graduate with surprisingly little amount of reading and writing, since most of our focus is on the mathematical side of things. Most of my studies were about formulating optimization problems and solving them mathematically. Typically the exams required no excessive amount of reading, whereas the problem sets and mathematical examples are important to learn. As a result, one can graduate with flying colors without being very impressive writer (or reader).
Anyway, I participated a scientific writing course, which turned out to be very helpful at this point. Only now have I really started to understand how important it is to actually learn to clearly express your ideas. I think that in mathematically oriented fields it’s rather common to ignore the importance of writing skills. One learns the true value of such skills just when one really needs to publish something. In addition to scientific writing course, I also attended a statistic modelling with R course. This course has also been a very informative one and I’ve learned a lot of important things. I think that statistical modelling is a next step from the basics of statistics. It has also been a very nice course for me because it takes a more practical approach as it focuses on examining the models through examples that are solved with R. In addition, the course is an online course and all the lectures are just in Internet. In addition to theses courses, we have the PhD level course “Economic growth and environmental resources”, which is basically a macroeconomics course. We have to hand in the exercise sets by the end of this month. There require a lot of attention and it’s hard work because I’m not familiar with macroeconomic models. Nevertheless, I’m familiar with the dynamic optimization, which appears to be a primary tool in solving the macroeconomic problems. Therefore the course is very useful for me, because I have to keep practicing all kinds of dynamic analyses. It’s also good to get basic understanding of the most important macro models although I wouldn’t ever do macroeconomic analysis within my work.
We also had a project meeting Stockholm. I gave a presentation of our paper there. I thought it went better than my previous presentations since my English speaking has improved quite a bit since I started and also because the work is now ready. Despite of those improvements, I was still very nervous and I talked very quickly so that no one could interrupt me. As a result, my presentation was too difficult to follow at that point of a day; I realized that I used only abbreviation of the soil phosphorus and I didn’t explain it at any point during the presentation. Thus, I may have lost the audience right in the beginning. Thus, the take-home-message from Stockholm is that I must learn to give clearer and simpler presentations in the future. This will be particularly important, because the paper of ours got accepted to EAERE (European annual environmental and resource economics) 2017 conference. This was very good news and it made me very happy. The conference will be held in Greece (I have never been so far a way from home before). Thus, it will be exciting both professionally and also as an adventure. I look forward to it. At the mean time, I must learn the art of giving presentations. In addition, I must make really rabid progress with the second paper, if I want to graduate in time. Thus, the spring and the summer will be busy, but also exciting time.
See you next time!
Best regards: Matti Sihvonen
20.02.2017 14:01One-way ticket from Matlab to R, and back!
Hi everybody!I’ve decided to write this somewhat monthly blog post in somewhat more technical fashion compared to my previous posts, which are written in very general fashion. The methods that I’ve been using so far in writing my first two articles have been non-linear weighted OLS estimation and dynamic programming. I’ve used Matlab for both of these. Matlab is a great program for many things, particularly optimizing, as far as my experiences are considered. However, I’ve started to question the reliability of the estimation results that Matlab provides. In addition, I haven’t been able to get all the statistics that I would need for the proper analysis. These statistics include the basics standard errors, t-values and p-values for the estimates. I also can’t reproduce the goodness-of-fit statistics by basics calculations myself. I noticed that the 95 % confidence intervals that Matlab gives for the estimated parameters are very wide. Those even go beyond the range were the function is determined, which I found very suspicious. Thus, I turned to R, which is very much-applied program for statistical analysis. I learned to do the weighted non-linear estimation in R and I finally got all the statistics that I needed. In addition, I understand the statistics it provides and there’s nothing mysterious going on, which I found very important when the actual research work is considered; one has to know where all the numbers have come from.
to say, I had to reconsider some models and results. However, although I found
R more suitable for statistical analysis than Matlab, the issue with non-linear
estimation in R is the starting values for the estimates. It seems that those
have to be very close to final values so that the program will be able to do the fitting. So
how do you get to those starting values? I used Matlab for those. Matlab’s
curve fitting toolbox, although I don’t trust in the statistics it provides, is
very useful, as it illustrates the curve and the residuals right away. It also
is very successful in fitting the curve or surface even with the default starting
values, once the iteration account is set high enough. This might be related to
Matlab’s good optimization abilities, as it is matrix-based program, and
estimation is optimization, after all. Thus, I used Matlab for initial
examination for a particular functional form and to get the starting values.
Then I inserted the functional form and the starting values into R to get the
final estimates for the parameters as well as the associated statistics. Then I
used Matlab again for drawing illustrative simulation figures and for the economic
optimization. I must say that for those purposes the Matlab is superior
compared to R, at least as far as my very limited programming skills are
considered. All the figures in my work are drawn with Matlab, with the
exception of one schematic diagram.
now that I’ve been shifting back and forth between Matlab and R a couple of
weeks, I believe that the first article is finally ready and it’s time to
continue the work with the second article, which focuses entirely on
optimization. In first work we
concentrated on model derivation and examination of the structural and
parameter uncertainty. In the next paper we will examine analytically and
numerically the private optimums a bit further and then we move to examine the
social optimum, where also the environmental externalities are considered. I
will write more about those in the upcoming blog posts. I guess my main tool when doing the second article will be Matlab again, because at least currently is seems that no statistical analysis will be involved.
Thus, see you next time!
Best regards: Matti Sihvonen
30.12.2016 16:58Discoveries of the year
I’m sorry that it’s been so long since my last blog post. The end of the year has been very busy. I’ve taken some courses concerning modelling, econometrics and statistics in order to learn more about the subject of my PhD. I’ve been so involved in these actions that I’ve haven’t got a time even to write this blog. I’m so sorry about that. Fortunately, the intense period is starting to be over now, at least for a while.
The courses that have kept me busy have been advanced econometrics, the extension course in introductory statistics and linking data and ecological models. All of these courses have been excellent and very useful; I’ve learned a great deal about things that are very relevant regarding my PhD. I must say that the course lecturers at the econometrics and linking data courses where exceptionally enthusiastic and inspirational. From linking data course I learned about the model comparison method called AIC, Akaike’s information criteria. I liked it so much that I used it to rank the models in the first paper. I also learned a lot about result reporting and validation methods.
From the advanced econometrics course I learned especially that I really have to examine the residuals for traces of heteroskedasticity or other distortions. I found some problems and we had to one more time to recheck the data. All kinds of transformations were tried, but as things are typically more complicated with the real data than with the book examples, we solved the problem by excluding some observations from the data set. These observations were related to some very old experiments and as such they came from the different distribution than the rest of the observations. Many things were discussed during the econometrics course and there was a lot of stuff to learn for the exam. Thus, I studied intensively for it. I’m happy to say that it went very well. Now I can finally focus entirely on finishing the first paper.
However, there is still a lot of writing to do before the first paper is finally ready. I’ve received comments from my supervisor Kari Hyytiäinen and co-author Elena Valkama. Before I could go off to my Christmas holidays I had to go through the comments and rewrite a lots of stuff as well as once again recalculate the results as the functional forms changed a little bit because of the residual distortions. Thus, there was a lot of work to do. There’s a light at the end of the tunnel; according to Elena, the paper is going to be a good one. That’s a very nice thing to keep in mind because this has been a very challenging task for me. The challenge originates from the fact that I didn’t have any background with statistics or econometrics or data analysis in general before this work, which is basically advanced biometrics, or something like that. Well, after this, I hope that the second and third papers will be easier. At least those are more related to economics, which I am more familiar with.
Currently I’m finishing the discussion section. I have this tool, which I use to organize this section. It’s kind of hard to understand what are the actual results in this kind of a paper, in which the model is built. Anyway, the paper is now almost ready and I hope that I have finished all the writing by the end of the Christmas holidays.
Thus, the battle rages on. The life of PhD student is rather heavy, I must say. At the same time, it’s very rewarding as we learn so much new things constantly. To wrap up the year a little bit I would say that the most important thing has been the discovery of the role of statistics in research work. I’ve always thought that the mathematics is the most important language in research work but this year I’ve understand the statistics is a very important language as well. Or actually statistics should be considered as a method whereas the mathematics is clearly a language, a way to present things accurately. In addition, I’ve discovered that the computer program R is something that must be learned. It is the way of the future. It almost seems that the statistics equals R these days. Well, this is very fine as the R is free to download for everyone and as such it’s a natural choice for universal environment for statistical programming. I must keep learning R because currently I know just the very basics.
I’ve also understood that as my own mathematical capabilities are limited, the statistics are even more important to me. This is because if I’m fortunate enough to continue my resource work in the future, I should probably focus more on empirical work; working with real data. And perhaps even if my resource career would be over after this PhD program, I could work with the data in some firm, or something. In any case, I feel that the empirical work including data analysis and statistical inference might be something that I could do in the future. At least it would be more my thing than highly theoretical and mathematical work, as I’m simply not talented enough in that field. It’s important to recognize and accept ones own limitations and strengths also.
Thus, back to books, happy New Year for every one, let the year 2017 be full of new discoveries.Best regards: Matti Sihvonen
03.10.2016 08:55Fuzzy business and basic statistics
In first week of September we had a PhD course, whose topic was fuzzy logics. That was a Matlab course, so I took it even though I knew nothing about fuzzy logics. Well, I still don’t know much about it. To me it seemed like a black box: you put some data into the computer and it builds some model based on some decision rules. The point seemed to be that the computer can build a better model that I could, but the computer won’t let me know explicitly what kind a model it has build. Instead, it will tell me how good the model is. We may check the structure of the model, i.e. how many rules it has. We can also check the surface of the model. The surface is usually much more complicated than in the case of a traditional mathematical model. This results of course from the fact that it’s difficult to build a mathematical model whose surface isn’t concave, convex or linear in a given dimension.
The course was interesting but I didn’t get
much out of it. Maybe I’m not advanced enough to understand the fuzzy
reasoning. I still have so much to learn about basic methods that I’m not in a
position yet to deal with methods beyond the basics. However, we still have an individual assignment to do for the course. I might learn more during the process. There are many applications for the fuzzy logic and fuzzy systems. After all, world is much fuzzier in reality as it's typically described in traditional modelling framework. Thus, the fuzzy reasoning something to focus on in the future research work.
Speaking of basics, as the work that I’m doing right now is very much about data analysis and statistics, I thought it might be a great idea to start studying statistics. Thus, I enrolled into a basic statistic course. In next period there will be a follow up course. This is also very good for me because my plan is to enroll into econometric course, which starts in next period. The econometric course is master level course and the statistics course is just basics. Next year I will take the PhD econometrics course. I assume that it would be impossible if I wouldn’t take these courses before it. I also hope that once I’m graduated and become a doctor, I will get some work, which will be related to empirical data analysis. Thus, the studies related to statistics and econometrics are essential to me considering my future plans related to my career.
Through these are just basic courses, they are very useful to me. One may always study the basics by oneself from any book, but when you’re actually taking a course, one is forced to do exercises and read more carefully. This is the case at least in my opinion.
However, I must confess that I have a little contradictory attitude towards the statistics because I’m very interested in empirical statistics and data analysis and all kind of related subjects but at the same time I know that among the different fields in mathematics, it’s the probability theory that is least appealing to me. I just can’t get my head around it and I never could. Even in high school I barely passed the probability course. And of course the statistical probability is in the heart of statistics. Now the first three exercise sets have focused on probability and last one didn’t actually go very well. The things you left behind can be found from in front of you, I see.. So it’s a bit of a struggle, but it’s a very beneficial one. These are the basics one just must know if one wants to do any kind of empirical work.
The actual work has been progressing quite well. I’m putting together the first article. We had a brief meeting in our campus with co-authors Elena Valkama and Eila Turtola and Elena seemed to be pleased with the work. She even suggested that I could make a review article about the functional forms typically used in yield response modelling articles. However, we might stick with the initial plan at least for now on. That is, by the end of the year, the first two articles must be ready and the third must be on its way. So, currently I’m finishing the first one, working hard with the material for the second one and reading for the third one. By the end of this month I must write a report about the subject of the third paper for the course that was held in Kristianstad this autumn. Thus, it will be a busy month and busy remaining year. I guess there is a lot a fuzziness ahead, but at least I know now how to throw it into the computer. What comes out, is another thing.
Best regards, Matti Sihvonen
02.09.2016 09:09To PhD is to travel – Roskilde and Kristianstad
Some may have noticed that I didn’t write about July. It’s because I took a little summer vacation to give my poor brain a rest. So there was nothing to write about in there. I could mention that I read some interesting political science fiction (Orwell, Huxley), some economic history and small amounts of biology and psychology. The weather in Finland was very nice. The summer was good and quite relaxing, charming even.
This August, in the other hand, was very intense. We had to produce a deliverable of our project by the end of the month. We had a two days meeting at Roskilde where we brainstormed about the scenarios and nutrient development path among other things.
I also held a presentation in the meeting about my model. Main objective was to illustrate how phosphorus stock develops over time and how the phosphorus leaching follows this stock development path. It turned out that my model really is useful for the project. This was a happy discovery for me because I was already in the belief that my work has separated from the project. However, some seemed to prefer not to use this kind of a model but take a shortcut instead. This comes down to the point where one has to bounder between simplicity and certain level of accuracy. There’s always a limited amount of data available. This is the case especially when one tries to examine big geographical area like Baltic Sea catchment on a long time horizon like almost 100 years. When working on such a wide scale it might not be practical to use detailed model. We may for example assume that nitrogen leaching follows directly the nitrogen input use. However, with phosphorus this is just not the case. This I tried to show in my presentation. So I was happy to notice that older and more experienced person verified my arguments when they were tried to ignore.
I also had to think how to calibrate the model for other countries. I used Eurostat data for fertilizer application rate and average yields. I know that this kind of calibrating is very coarse and the accurate thing to do would be to estimate own response, transition and leaching functions to every region. Calibration was straightforward and practical way to proceed within this time limit. After the calibration August was very much about running the model and producing results for scenarios.
In addition to Roskilde I also participated in PhD course in Kristianstad, Sweden, which was organized by Miracle. The course included topics like social learning process as a governance tool, HYPE modelling, flood control, common agriculture policy (CAP), practical field work like gathering water quality samples and cost benefit analysis. The lectures were quite good and I think my thinking may have expanded a little bit. Especially the lecture about CAP was very informative and useful for me because we’re dealing with EU level agricultural policy measures in our project. I also started to see a role of a scientist a little bit differently. A role of modelling as a part of decision making was also discussed, which I found interesting. Most valuable thing, however, was meeting people and communication in English for whole week. We had very good conversations and I met very nice and interesting minds. I hope I will meet some of them in the future.
However, not all went smoothly during the week. We were divided into groups in the beginning of the week and we had group exercises every day. In addition we had to write a report and held a presentation in end of the week. This would have been very nice but the communication and working dynamics within our group just didn’t work. I don’t know whether this resulted from cultural differences or personal problems or some other aspect. Anyway, it was most difficult group work situation that I have ever been in. I was absolutely amazed that we could pull the presentation together in the last minute and actually got through it without a complete breakdown. This was a bit funny situation, considering that we are already PhD students. Maybe I was the problem. I am used to work all day and after that to study in the evenings. Not all are used to this kind of way to work. But I didn’t push the others. It would have been perfectly fine even if I had done the whole thing. Well, anyway, sometimes thing just don’t work with people. Luckily there’s still an individual assignment that have to be done for the course. This has to be related directly to one’s PhD work. Thus, I relate it to my third paper, which handles risk and uncertainty within farm business. This is a very good opportunity to push that work forward. My ambitious goal is to get my first two papers ready by the end of the year and at the same time start to write my third paper. If this would succeed I would be in schedule.
Let the show go on and see you in next month!Best regards, Matti Sihvonen
06.07.2016 10:27Insights about modeling
During the last month I’ve learned a great deal about the fundamental aspects of modeling. I thought that work for the first paper was almost finished. So I started to seek information and write about the theoretical aspects of modeling. During the process I found such a good references that it actually helped me to figure out some totally new functional forms that I haven’t thought before. It didn’t take long until I got 20 functional forms for the model. I looked for reading which would help me to determine which functional forms to apply in the final model. I learned that there are at least three fundamental aspects that one needs to consider. First, one needs to think carefully what is the phenomenon that one is trying to model. That is, what assumptions one have to make about the properties of the model. Second, one takes a careful look at the data and examines it as deeply as possible. Third, one have to consider what for the model is to be built. That is, in what applications the model is to be used. The model have to work in those applications.
My supervisor Kari Hyytiäinen told me that model builders could basically be divided into process based and data based people. It seems that economist tend to be often process based whereas ecologist tend to be more data based. This takes us to the resent problems I’ve been struggling with. When we started this modeling process we made some assumptions about the properties of the model. This was before we even looked at any data. Then we contacted ecologists and got involved with the great data. We got into some arguments with the ecologists because they wouldn’t accept the assumptions we made because, although the assumptions were reasonable, the data didn’t support those.
So I jumped into the data based train for a moments; I removed the theoretical element related to the yield increasing effect of the soil phosphorus (STP) from the model. The ecologists were happy and I managed to build a model that delivered the aspects that were observable in the data. Then I started to focus on the application of the model: optimization and simulations. That’s where I run into trouble again. First I noticed that there was something suspicious going on with the optimization process: why the optimal soil phosphorus path didn’t go to zero if it was only bad in the model? I also noticed that absolute amount of the yield decreased as the STP increased when the model was simulated. This happened because STP was the bad guy in the model, but this was highly unintuitive and it wouldn’t make sense in the real life. I realized that the problems related to the missing model element describing the response to STP and the transition function describing the soil phosphorus dynamics.
At this point I had a meeting with a great environmental economist, Antti Iho, who knows a great deal about phosphorus modeling, among other things. He also noticed what was wrong. We realized that the transition function, which was suggested by the ecologists, was a failure. It only worked on a very limited data range. This is the problem with highly data based modeling; the produced model makes sense on a limited domain and thus it is not useful for any applications. I don’t think it’s a good approach. Anyway, Iho also said to me that we just couldn’t remove the increasing element of STP; the simulation and optimization result will be biased. This was true of course. It only confirmed my suspicions; the initial assumptions of us were correct. The problems won’t disappear if we just look the other way for a moment.
Hence, here I go again, in the conflict of the two modeling camps. But now I know, that the tree fundamental aspects of modelling must hold simultaneously. Otherwise the model just isn’t good. Currently I try to overcome this challenge, related to the mystery guy, STP, by applying additional data from the great report by Saarela (1995). Hope the model will be ready soon, before I completely loose my head with it.
Best regards: Matti Sihvonen
25.04.2016 09:48Scenarios, asymmetric information and rocket science
22.12.2015 15:34Nightmares before Christmas
This is my last blog post this year. It has been quite a demanding December (and late November). We had a very difficult final exam in the microeconomic course. I really hope I passed it. I’m waiting for the results with fearful thoughts.
03.12.2015 11:35Struggle with the fertilizers
Hi everybody!This August has been all about the fertilizers for me. As the aim of this work is to investigate the cost efficiency of the abatement measures in the long run, we have to form scenarios that describe the future development of the Baltic Sea region.
03.08.2015 10:27Summer at university
July was a very quiet month here at the Department of Economics and Management in Helsinki University. All the staff were enjoying their summer holidays. However, I spend the whole month here working and studying. It didn’t even feel that bad because of the weather, which was rather awful throughout the whole month.