News on China's scientific and technological development.

Icmer said:

Okay, here is the comparison that you claim, it seems, to be solely necessary.



Does this tell you anything about efficiency of resources allocated? Not at all.

Let's look at share of researchers again.

Please, Log in or Register to view URLs content!

Compared to China, South Korea has a much higher proportion of researchers in the workforce (14.43 per 1,000 people) than China (2.242) despite South Korea's R&D spending being 4.3% of GDP compared to China's at 1.97%. Similarly, the most recent figure for the US is 8.928, while US R&D as share of GDP is 2.84%.

Do you deny that this discrepancy is at all notable? This is exactly the same as what the original article discusses, except in a more distributed form.

Click to expand...

That article was not even particularly looking at the "efficiency" of Chinese R&D, but rather looking at overall Chinese R&D progress. For that purpose, yes, looking only at the absolute amount of R&D expenditure and the absolute number of R&D researchers can be useful.


Furthermore, if we were interested in looking at "efficiency" of resources allocated then we would be looking at something different -- namely the absolute amount of R&D budget, and the ratio to absolute no. of R&D researchers!

If you were interested in looking at "efficiency," well you still need the absolute amount of R&D expenditure as well as the absolute number of R&D researchers anyway.



Now, I'm not saying that what I described briefly here in these couple of posts are able to perfectly capture R&D progress or R&D "efficiency". They're just measures I've come up with off the top of my head.
However, I am definitely saying that the variables they've chosen to use are very illogical and far, far less useful.

Bltizo said:

That article was not even particularly looking at the "efficiency" of Chinese R&D, but rather looking at overall Chinese R&D progress. For that purpose, yes, looking only at the absolute amount of R&D expenditure and the absolute number of R&D researchers can be useful.

Click to expand...

Taken at face-value, its utility is limited. Hence why it is important to deconstruct different aspects of Chinese innovation. We're only discussing "efficiency" because it is what the section in question is about. You can see the other variables used through the rest of the article. Would you argue that the % share of citations in top journals relative to the US is not a good metric of "overall Chinese R&D progress"?

Furthermore, if we were interested in looking at "efficiency" of resources allocated then we would be looking at something different -- namely the absolute amount of R&D budget, and the ratio to absolute no. of R&D researchers!

Click to expand...

For these comparisons, these absolute amounts were already employed. In statistical analysis, a more refined metric is usually necessary, or else your analysis would be lacking in rigor and precision.

If you were interested in looking at "efficiency," well you still need the absolute amount of R&D expenditure as well as the absolute number of R&D researchers anyway.

Click to expand...

It should be obvious that the authors did indeed use these figures to create such metrics.

Now, I'm not saying that what I described briefly here in these couple of posts are able to perfectly capture R&D progress or R&D "efficiency". They're just measures I've come up with off the top of my head.
However, I am definitely saying that the variables they've chosen to use are very illogical and far, far less useful.

Click to expand...

They're not illogical if you understand the usefulness of having simple, direct comparisons. And I see that you're no longer contesting the usefulness of the original comparison after I explained it to you.

Icmer said:

Taken at face-value, its utility is limited. Hence why it is important to deconstruct other aspects of Chinese innovation. You can see the other variables used through the rest of the article. Would you deny that the % share of citations in top journals relative to the US is not a good metric of "overall Chinese R&D progress"?

Click to expand...

You and I are arguing about different things.

You're arguing that the variables I've given are not the best measures for capturing R&D or innovation -- I fully agree with you, and I've never suggested anything contrary to that idea.

However, my argument is that their variables ("share of XYZ relative to US") are far, far worse and near useless for the purposes of comparison that they're trying to do.

For these comparisons, these absolute amounts were already employed. In statistical analysis, a more refined metric is usually necessary, or else your analysis would be lacking in rigor and precision.

Click to expand...

I agree that more rigorous analysis is beneficial and I am very interested in investigations in that domain.
However comparing "expenditure as share of GDP relative to US" and "R&D workers share of overall workforce relative to US" is not "rigorous analysis". It is a red herring at best, and idiotic at worst.

It should be obvious that the authors did indeed use such figures to create these metrics.

Click to expand...

The authors did not mention absolute numbers for their "expenditure as share of GDP relative to US" nor did they mention absolute numbers for their "R&D workers share of overall workforce relative to US".

My entire argument is that their statistics are useless without giving us the absolute numbers.

They're not illogical if you understand the necessity of having simple, direct comparisons.

Click to expand...

Their comparison is not simple nor is it direct.
It is convoluted and inconsistent.

A simple and direct comparison would be comparing something like the ratio of absolute R&D budget vs absolute no. of R&D researchers vs absolute no. of citations in top journals (or something like that).
Comparing a ratio of those three variables between China and the US over time would be simple, direct and useful.

Comparing "R&D expenditure as share of GDP relative to US" with "R&D workers share of overall workforce relative to US" without even giving us the absolute number to begin with is just nonsensical.

Bltizo said:

You and I are arguing about different things.

You're arguing that the variables I've given are not the best measures for capturing R&D or innovation -- I fully agree with you, and I've never suggested anything contrary to that idea.

However, my argument is that their variables ("share of XYZ relative to US") are far, far worse and near useless for the purposes of comparison that they're trying to do.





I agree that more rigorous analysis is beneficial and I am very interested in investigations in that domain.
However comparing "expenditure as share of GDP relative to US" and "R&D workers share of overall workforce relative to US" is not "rigorous analysis". It is a red herring at best, and idiotic at worst.




The authors did not mention absolute numbers for their "expenditure as share of GDP relative to US" nor did they mention absolute numbers for their "R&D workers share of overall workforce relative to US".

My entire argument is that their statistics are useless without giving us the absolute numbers.





Their comparison is not simple nor is it direct.
It is convoluted and inconsistent.

A simple and direct comparison would be comparing something like the ratio of absolute R&D budget vs absolute no. of R&D researchers vs absolute no. of citations in top journals (or something like that).
Comparing a ratio of those three variables between China and the US over time would be simple, direct and useful.

Comparing "R&D expenditure as share of GDP relative to US" with "R&D workers share of overall workforce relative to US" without even giving us the absolute number to begin with is just nonsensical.

Click to expand...

These variables were directly taken from the OECD, as noted in the cited sources. In this case, according to academic style, it's completely unnecessary to have to repeat the raw data, since these metrics are literally the same as presented on the OECD website.

For example, Figure 1: "Chinese Expenditures on R&D as a Share of GDP, Relative to the United States, 2007–2017" is thus cited:

46. OECD, “OECD Stats: Gross Domestic Expenditure on R&D by Sector of Performance and Source of Funds,” accessed March 4, 2019,

Please, Log in or Register to view URLs content!

.

Click to expand...

Indeed, this metric can easily be found by accessing "Science, Technology and Patents" -> "Research and Development Statistics" -> "Expenditure".

It seems your argument is that the practice of applying the data as a proportion is too complex, and that such a simple statistical analysis requires the raw data to be tabled and presented in their totality. This is unnecessary if you have a reasonable amount of familiarity with how to read academic papers and verify sources.

Icmer said:

These variables were directly taken from the OECD, as noted in the cited sources. In this case, according to academic style, it's completely unnecessary to have to repeat the raw data, since these these metrics literally the same as presented on the OECD website.

For example, Figure 1: "Chinese Expenditures on R&D as a Share of GDP, Relative to the United States, 2007–2017" is thus cited:


Indeed, this metric can easily be found by accessing "Science, Technology and Patents" -> "Research and Development Statistics" -> "Expenditure".

Click to expand...

Yes, as I wrote in my last post, "It is a red herring at best, and idiotic at worst."

I'm not sure what was going through their heads when they thought it would be a useful to calculate irrelevant variables like "expenditure as share of GDP relative to US" and "R&D workers share of overall workforce relative to US".

Those numbers do not offer us anything more useful than repeating the raw data. If anything they are more confusing.

If they were trying to compare R&D output or potential then they could've compared absolute R&D expenditure of China vs US, or absolute no. of R&D workers in China vs US.
If they wanted to compare R&D "efficiency" then they could've used those variables I described above and compared them as ratios with output of citations in top journals or something like that.



I want to clarify -- I applaud any effort to try and better gauge R&D potential, R&D output and R&D efficiency.

However, the variables "expenditure as share of GDP relative to US" and "R&D workers share of overall workforce relative to US" alone are almost completely useless to us in gauging those outcomes. They are confusing at best and distracting at worst.

Bltizo said:

If they were trying to compare R&D output or potential then they could've compared absolute R&D expenditure of China vs US, or absolute no. of R&D workers in China vs US.
If they wanted to compare R&D "efficiency" then they could've used those variables I described above and compared them as ratios with output of citations in top journals or something like that.

Click to expand...

Apparently you didn't acknowledge that those figures are quite limited in scope and alone do not lend themselves to much analysis. A short news piece would be sufficient coverage of those numbers, such as this

As for journal output/citations, I can see how the metrics in Figure 12 can be combined with expenditure. But it would still leave open the question of how there is so much R&D spending when there are proportionally so few Chinese researchers compared to other countries.

I want to clarify -- I applaud any effort to try and better gauge R&D potential, R&D output and R&D efficiency.

However, the variables "expenditure as share of GDP relative to US" and "R&D workers share of overall workforce relative to US" alone are almost completely useless to us in gauging those outcomes. They are confusing at best and distracting at worst.

Click to expand...

So I see you just dislike those variables because you find the "relative to the US" part confusing. I don't think you would object if the data were presented in the format I brought up earlier, even though they are practically the same. Unless you don't see the utility of discussing that discrepancy altogether. By the way, this is hardly the first article to mention it. I'll link another article I remember reading if I can find it.

Icmer said:

Apparently you didn't acknowledge that those figures are quite limited in scope and alone do not lend themselves to much analysis. A short news piece would be sufficient coverage of those numbers, such as this

Please, Log in or Register to view URLs content!

Click to expand...

Of course those figures would be limited in scope.
However at least they wouldn't be flawed.

As for journal output/citations, I can see how the metrics in Figure 12 can be combined with expenditure. But it would still leave open the question of how there is so much R&D spending when there are proportionally so few Chinese researchers compared to other countries.

Click to expand...

Can you show me the numbers for how Chinese R&D spending is high relative to its R&D workforce?

We need absolute numbers of Chinese R&D spending and the absolute number of China's R&D workforce to ascertain what the relationship is.

Comparing the "share of GDP" and "share of overall workforce" are not viable proxies for trying to extrapolate how "reasonable" the relationship between China's R&D spending and China's R&D workforce is.

Basically, what you are asking about is what the R&D spending per R&D workforce capita is. Putting it another way, you're asking about dollars spent per R&D worker.
However, for that you need absolute numbers, which the article doesn't provide.

So I see you just dislike those variables because you find the "relative to the US" part confusing. I don't think you would object if the data were presented in the format I brought up earlier, even though they are practically the same. Unless you don't see the utility of discussing that discrepancy altogether.

Click to expand...

No, I dislike these variables because I find the "share of GDP" and "share of workforce" parts unnecessary. The "relative to US" part is useless as well, but the big problem is the "share of GDP" and "share of workforce" part.

Let me give you a simple example for why their variables are ridiculous. I'm going to use vastly exaggerated scenarios below, but it will serve to demonstrate my point for why their chosen variables are stupid.

Let's say I am a country -- The United Republic of Blitzo, population 1, me. Let's say that my annual GDP is $10,000 and I invest $5,000 of that into my full time job is as a researcher (my annual R&D budget)
Based on that, my "R&D expenditure as share of GDP" is 50% ($5,000/$10,000)
My "R&D worker as share of overall workforce" is 100% (1 person/1 person).

Let's compare that with country ABC. Country ABC has a population of 100 million. Let's say its annual GDP is 2 trillion dollars. Let's say it invests 40 billion dollars into R&D per year. Let's say that its working age population (overall work force) is 70 million, and let's say out of its 70 million strong workforce 700,000 of those work in R&D.
Based on that, Country ABC's "R&D expenditure as share of GDP" is 2% ($40 billion/$2 trillion)
Country ABC's "R&D worker as share of overall workforce" is 1% (700,000 people/70 million people)

Let's recap those numbers:
The United Republic of Blitzo has an "R&D expenditure as share of GDP" of 50%, vs Country ABC having an "R&D expenditure as share of GDP" of 2%
The United Republic of Blitzo has a "R&D worker as share of overall workforce" of 100%, vs Country ABC having "R&D worker as share of overall workforce" of 1%.

So, if those were the only statistics you had (underlined in the paragraph above), how would you interpret those numbers? Are they useful for trying to derive anything about either country's R&D potential, output, or efficiency?

Obviously not! Because they don't tell us anything about the absolute GDP size of the URB vs Country ABC, and they don't tell us anything about the population of the absolute workforce of URB vs Country ABC.


Instead, what I should've done is to tell you that my upstart Republic has a budget of $5,000 vs Country ABC's $40 billion, and that my R&D workforce is 1 person vs Country ABC having 700,000 people!

The same principles apply when looking at real life nations like China and the US.


By having the absolute numbers to begin with, we are able to try and compare those numbers with other additional factors if one is interested in looking into things like R&D output efficiency in more detail.

But the problem is that the article only gave us "R&D expenditure as share of GDP" and "R&D worker number as share of overall workforce" without giving us the absolute numbers for ourselves to try and work with.
Instead, what they've done is make things more complicated and idiotic because they're making it seem like their variables are useful, when in reality it's the exact opposite and if anything they are misleading without context (which I want to repeat -- they did not give us context!)

Bltizo said:

Instead, what they've done is make things more complicated and idiotic because they're making it seem like their variables are useful, when in reality it's the exact opposite and if anything they are misleading without context (which I want to repeat -- they did not give us context!)

Click to expand...

Actually they did.
These indicators are purposely narrow so that they would be easy to directly track over time. Taken altogether, think of them as a distinct form of Li Keqiang Index.

ASSESSING CHINESE INNOVATION PROGRESS

A thorough examination and answer to the question of current and future Chinese innovation capabilities is beyond the scope of this report. However, by examining 36 indicators of China’s scientific and technological progress vis-à-vis the United States a decade ago versus today, it is possible to get a sense of where China is making the most progress, and to what extent it is closing the innovation gap with the United States. Indeed, If China is only a copier, then the risk to advanced economies is limited. But if China is more like the “Asian tigers” that rapidly evolved from copiers to innovators, the threat is serious.

INFORMATION TECHNOLOGY & INNOVATION FOUNDATION

China has made progress on all indicators, and in some areas it now leads the United States. In fact, in an average of all the indicators, China has cut the gap to the United States by a factor of 1.5 from the base year to the most recent year. (For example, had China been 80 percent behind the United States a decade or so ago, it would be just 50 percent behind in the most recent year.)1 In other words, in the span of about a decade, China has made dramatic progress in innovation relative to the United States. This report briefly highlights Chinese government goals and policies to become a world innovation leader. It then reviews the claims skeptics make about China’s ability to innovate—arguing that, for the most part, their definition of the term “innovation” is too narrow—and reviews arguments and evidence supporting the claim that China is successfully innovating. Finally, the report presents the 36 indicators.

Click to expand...

CHINESE INNOVATION INDICATORS

Actually measuring China’s innovation capacity and performance relative to the United States is difficult. Ideal indicators include innovation performance of Chinese and U.S. firms, but that data is not collected. There are, however, a variety of indicators that are available, which can be sorted into three categories: 1) inputs into the innovation process, such as R&D spending and scientific and engineering personnel; 2) outputs from the innovation process, including patents and scientific articles; and 3) outcomes related to actual innovation, including industry sales and exports. To be sure, economies and companies differ significantly in how effectively they translate inputs—even outputs into actual innovation outcomes. The Chinese innovation “machine” appears to be significantly less efficient than the U.S. system in translating inputs to outcomes. Nevertheless, inputs and outcomes matter to innovation success. The goal of using these indicators is not to come up with a definitive measure of how innovative the Chinese economy is relative to the U.S. economy, but rather to explore Chinese progress relative to U.S. progress on a host of innovation indicators. This report uses 36 indicators to do so, measuring Chinese performance relative to the United States in the most recent year for which data is available, and a base year (usually ten years prior) to assess the rate of Chinese catch-up. For indicators of inputs and outputs, it controls for the relative size of the economies. For most indicators, Hong Kong is included as part of China. On all the indicators China has closed the gap or, in some cases, extended its lead over the United States. In fact, in an average of all the indicators, China has cut the gap to the United States by a factor of 1.5 between the base year and the most recent year. For example, had China been 80 percent behind the United States a decade or so ago, it would be just 50 percent behind in the most recent year.

Click to expand...

As for figures such as the total GDP, total workforce, and total expenditure, I argue that repeating them is unnecessary. Only laymen would have difficulty finding this information, and the audience of this article would certainly understand the scales involved. If they're really worth mentioning, they would be best discussed in the introduction, as a brief overview.

Icmer said:

Actually they did.
These indicators are purposely narrow so that they would be easy to directly track over time. Taken altogether, think of them as a distinct form of Li Keqiang Index.

Click to expand...

... First of all, when I say they did not give us context -- I mean absolute numbers. I.e. if they want us to make sense of their "share of XYZ" variables, they did not give us the context to make sense of it because the necessary context to make sense of it requires absolute numbers to give us meaningful statistics that can be compared on equal footing.

Whether it is taken over time or not is irrelevant to the fact that the variables they chose to calculate have no statistical or logical reason to be considered without giving us absolute numbers.

Please re-read my previous post and carefully consider my United Republic of Blitzo Vs country ABC example for why the variables are useless.

Bltizo said:

... First of all, when I say they did not give us context -- I mean absolute numbers. I.e. if they want us to make sense of their "share of XYZ" variables, they did not give us the context to make sense of it because the necessary context to make sense of it requires absolute numbers to give us meaningful statistics that can be compared on equal footing.

Whether it is taken over time or not is irrelevant to the fact that the variables they chose to calculate have no statistical or logical reason to be considered without giving us absolute numbers.

Please re-read my previous post and carefully consider my United Republic of Blitzo Vs country ABC example for why the variables are useless.

Click to expand...

Like Zeke, you're again interpreting the variables in a vacuum. An individual section discussing a single indicator does not need to repeat China's GDP or that its workforce is hundreds of millions larger than that of the US. That should be self-evident.

Look at the indicator at the center of this debate again. Figure 6: Chinese Researchers as a Share of Total Workforce, Relative to the United States, 2006–2016. If it rises from the present 25% to 50% in 10 years, that is clearly a win for Chinese higher education and potential research output, and would mean China is further catching up to the US in this area.

Of course the US is taken a barometer for these variables. It is presumed that the US will consistently be the leading country globally in the near future in sci-tech. It's unlikely that there will be dramatic shifts in either country's position (China vs. US) in terms of GDP, expenditure, workforce, etc.