Saturday, 8 September 2012

Caterpillar: Productivity in Recession

Caterpillar: Productivity in Recession: I recently heard a pod cast from Econtalk, ‘Ohanian on the Great Recession and the Labor Market’. The host is Russ Roberts and the guest i...

Productivity in Recession


I recently heard a pod cast from Econtalk, ‘Ohanian on the Great Recession and the Labor Market’. The host is Russ Roberts and the guest is Lee Ohanian. The related pod cast link is here: http://www.econtalk.org/archives/2012/08/ohanian_on_the.html. In this podcast, it mentioned that in the past recessions before this GFC, the US’s productivity and employment both decreased, whereas this GFC makes the US experience increasing productivity and decreasing employment. They didn’t get into too much detail about this subject, which I was pondering for a while. Now I am going to share my personal theory on this subject with you.

Let’s put this GFC aside and focus on the past recessions for a moment. In a typical economic recession, market demand in some/all industries is close to saturation. Under such circumstances, business has a number of options:
1.  Finding a new business opportunity in a new industry
2. Introducing a new product/technology to the existing business
3. Decreasing the price of the existing products and doing clearance sales, considering the relative storage and maintenance cost
4. Conducting retrenchment and laying off employees

The first two options are intangible-asset-related work. It is very hard to measure the productivity in this sort of work. There is a great deal of exploration/ trial-and-error work involved.  Moreover, assuming the new technology is existent already, it is also hard to measure the productivity of adding business value to this new technology. When I say adding business value to a new patent/technology, I am referring to:
1.  Making this technology economical viable
2. Making the customers understand the value of this new-technology-related product
3. Creating the legislation related to this product
4. Making the industries related to this new technology economical viable

I would like to explain the last point in more detail. Let’s say you develop a new machine that makes metal faster and better. The related industries to this new machine may include:
a.       The manufacturing factories that produce this machine and may need to upgrade their system to produce it
b.      The training programs that train workers to operate this machine and/or operate the upgraded system mentioned in Point a
c.       The electricity supplier that may need to upgrade their machine to adapt to this new machine and so on

All of these are intangible-asset-related work and it’s not easy to measure the productivity of this work.  Though I haven’t had a chance to have a look at the productivity data mentioned in the podcast, I believe neglecting this intangible-asset-related work will make productivity underestimated, as it is hard to measure it especially when the business value of this intangible-asset-related work is not realised yet.

As I said before, the business can also reduce the price and do sales on their existing products. This may suggest that the same number of sales people may sell fewer products in recession given the market demand is weak. This also means that the productivity decreases.

In light of retrenchment, the story is not as purely simple as it seems. Although layoff means that the labour market is weak, from the productivity point of view, however, the answer is complicated. In theory, workers without sufficient training or updated knowledge will not get hired, whilst the new technology brings in better products in a more efficient manner. Empirically, the process of retrenchment and restructure is a process of political operation, which may include:
1.       Negotiation with the union
2.       Establishing new business network if the new business opportunity/ technology requires relocation across countries/industries
3.       Rebalancing the political power given the new source introduced by the new technology getting into the business

None of these are tangible-asset-related work. How can it be measured in terms of input versus output? What I am trying to explain here is that there is more intangible-asset-related work in recession than that in booming economy. Neglecting this makes productivity underestimated. If it’s true that the US’s productivity doesn’t decrease in this recession, how can you explain it? My opinion is that the US is going off the recession. They are realising or have realised the business value of this intangible-asset-related work already, so it can be seen that their productivity increases and employment is still an issue as they are too good in getting work done to hire more people.

In the end, I would like to thank Eugene Dubossarsky for listening to my idea and encouraging me to post it to my blog.  




Wednesday, 9 May 2012

Caterpillar: Book Review – Learning Marketing with Social Media...

Caterpillar: Book Review – Learning Marketing with Social Media...: Recently I read a book twice, ‘Learn Marketing with Social Media in 7 Days’ by Linda Coles. It’s a very good book and tells the reader how...

Book Review – Learning Marketing with Social Media in 7 days


Recently I read a book twice, ‘Learn Marketing with Social Media in 7 Days’ by Linda Coles. It’s a very good book and tells the reader how to utilise social media, such as Facebook, LinkedIn, and YouTube, to do marketing.

I am going to rephrase my favourite part of this book and share it with you. Social media offers a random person an opportunity of doing things online that he wouldn’t dream of doing in person. This also explains the flood of tweets and status updates. To stand out from the crowd, he needs to develop real relationships, which requires socialising skills/ people skills. In other words, how well his connections are, how well his product/service is presented, and how well the marketing program is going don’t depend on the social media, but the people who utilise the social media.

This reminds me of the article I read a few weeks ago, ‘Putting People First’. This article said and I quote, ‘Analytic tools are only as good as the people who use them’. In terms of Coles’ book, I would like to say that social media tools are only as good as the people who use them, constantly putting themselves back on the radar of their connections.

Saturday, 5 May 2012

Caterpillar: How do you learn R?

Caterpillar: How do you learn R?: Now I run a self-study group in R. The question I often come across is ‘How do you learn R? How much time did you spend learning this?’ Sinc...

How do you learn R?

Now I run a self-study group in R. The question I often come across is ‘How do you learn R? How much time did you spend learning this?’ Since then, I would like to share my learning experience. Please bear in mind that different people learn the same thing in different ways. My way of learning R might be different from yours, but if you can get some useful tips from my learning experience, I would be delighted to see the growth of R users.

R is a steep learning curve. It’s fairly easy to get started by using R manual. However, when you want to learn different R packages, you may find it difficult. A friend of mine once told me, 'there are no difficult questions, only poorly defined questions.' I agree. If you find it difficult to learn a specific R package, there are two questions that you need to ask yourself:

  • Do I understand the algorithms of this package or this function?
  • Do I understand the coding of this function?

Once you complete the R manual, I would not say you will have difficulties in understanding most of R coding. However, R manual wouldn’t be able to include all sorts of algorithms that are applied to all packages in R. This suggests that finding it difficult to learn a particular R package possibly means that you have difficulties in understanding the underlying algorithms. In other words, if you don’t know when you should use a particular function, you have problems with algorithms instead of R itself. This is where the package manual comes in. In the manual of each package, it introduces the algorithms it uses, providing leads that you can use to search in Wikipedia or YouTube or Google. Thereafter, you can find the relevant sources for algorithms to learn.

The package manual has the other function that it tells you what function does what, what parameter is required for what function, what parameter indicates what, and etc. Moreover, you can also use the examples in the manual to see how each function works. If you learn by using, the package manual is there for you. Also, ‘help()’ and ‘?’ are very useful. They extract the relevant piece from the manual and give your examples.

As I said before, different people learn in different ways. If you feel more efficient learning alone, you can do what I just said. If you feel more efficient learning in a community, you also have other options. R is open-source software and has a large user-community. I often go to R presentations, read r-bloggers and ask questions in a discussion forum if I have any. I also share my learning experience with other R users and we all learn from each other.

The last thing I would like to say is that when you learn R, you need to focus on the big picture instead of details. Manual is there to solve the details, whereas your mind needs to pay attention to the big picture, which may include:
  • What a package does
  • Under what circumstances you should use what function
  • What the underlying algorithm is


If you feel you get lost or difficult to keep learning, then you need to stop and step back to review what you learned and what is there in your big picture. Get the big picture before you get into details. In the end, thank Eugene for proof-reading and assisting me with running this self-study group.

Tuesday, 17 April 2012