24 July 2015


Dear Ben,

I read your Business World column of 7 July 2015  where you raise some good points. Unfortunately, the piece was marred by errors in some of the statistics and computations you use. I’m sorry, but I think we need to correct them for pedagogical reasons.

You write:

In 2010, there were some 7.3 million unemployed (completely idle) and 18.8 million underemployed (working but unpaid, working part-time, and working but actively looking for a better job). That’s 26.1% of the total work force or 10,158,000 warm bodies.

Actually, “7.3 million” is not the number of unemployed people; rather 7.3% was the unemployment rate for that year. (Look at the official figures here.) With a labour force of about 39.9 million in 2010, the total number of unemployed was 2.84 million.

Similarly, “18.8 million” is not the count of the underemployed. Rather you were probably referring to 18.7%, which was the underemployment rate for the year, or something like 6.75 million underemployed persons.

Then you say that that the number of the underemployed plus the unemployed equals “26.1 % of the total work force”. Unfortunately that is also wrong.

If one adds the correct number of the unemployed (2.84 million) and the underemployed (6.75 million), one gets 9.59 million or 24.6% of the labour force—versus the 10.16 million and 26.1% reported in your column.

The same miscalculation is made when you processed the figures for April 2015, about which you wrote:

 In April 2015, despite the slightly lower unemployment (6.4%) and underemployment (17.8%) rates, 24.2% of a much bigger work force, or 10,131,000 warm bodies, remain unemployed or underemployed. The number could be higher if workers who went abroad and workers in Leyte were counted.

Here there are two errors: the first was adding the unemployment and underemployment rates, i.e., 6.4 plus 17.8 = 24.2. That’s wrong because the base of the unemployment rate is the labour force, while the base of the underemployment rate is the employed.

If u is the unemployment rate and d is the underemployment rate, then the head count of the unemployed and underemployed taken together as a proportion of the labour force is not u + d but rather u + d(1 – u). I wrote a blog about this sometime ago. So in this case, the number is not 24.2 % but rather 23.1%.

This naturally leads to a second error. Since the labour force in April 2015 was 41.86 million, 23.1% of that gives you 9.67 million as being either unemployed or underemployed—not the 10.13 million stated in your column.

A third point relates to your use of only a single round of the labour force survey for 2015 to compare with the full-year results for 2014 (a result of four surveys). These two are, of course, not directly comparable, since there is still the performance for the rest of 2015 to be monitored, which can change the whole picture.

Just to show how things can change, we can (and in fact we should) use both the January 2015 and the April 2015 results by averaging them. Averaging, after all, is how full-year figures are ultimately arrived at, e.g., the labour force survey results for the full year 2015 will essentially be the average of the numbers for the January, April, July, and October surveys.

If we do this for the two surveys conducted in 2015, for example, we get a labour force of 40.99 million (average of 40.09 mn and 41.88 mn); average unemployment of 2.66 million (average of 2.68 and 2.65 mn), and so on. The average number of underemployed is 6.76 million, or a total of 9.42 million unemployed and underemployed.

To sum up, the proper figures to appreciate may be seen in the following table:









%   of




























*Labour force figures for January and April 2015 were 40.09 mn and 41.88 mn, respectively.

If one now compares 2010 not just with April 2015 but with Jan-April  2015, (referring to the table above) one sees that:

  • rates of both unemployment and underemployment have decreased relative to 2010;
  • the number of unemployed per period is lower in absolute numbers by some 180,000;
  • the number of underemployed per period however is higher by about 10,000;
  • all in all, therefore, the average number of people who were either unemployed or underemployed is lower by 170,000 in 2015 than in 2014; and
  • the overall rate of labour underutilisation (unemployed plus underemployed as a percentage of the labour force) has fallen from 24.6 % in 2010 to 23.0 %.

From this viewpoint, the figures do not seem as dire as your column paints, since labour underutilisation has fallen in terms of both rates and absolute numbers. (This is far from saying we should be satisfied, and  I agree that more can and should be done to improve jobs-generation — this is something the government itself admits.) For this reason, it will seem to some that your conclusions are unwarranted — or at least premature.

The larger issue, of course, as I already pointed out elsewhere, is that neither unemployment nor underemployment adequately measures welfare; that’s because most of the poor are not unemployed, nor are the unemployed mostly poor. The bulk of the poor are already fully employed and their problem is not work per se but low productivity and incomes.

There’s one last point. In your column you surmise that including overseas workers in the count would worsen the unemployment picture:

In its definition of the labor force, the Labor Force Survey (LFS) excludes overseas Filipino workers in the estimation of the size of the working population — that is, population aged 15 years and older. This means that the joblessness picture would have looked more serious if overseas employment were included in the work force.

This cannot be true, since if one includes overseas workers in the labour force (the denominator), then one must also include them in the numerator as being among the employed. If N represents the employed and L is the labour force, then N/L is the employment rate. Adding the overseas workers w, who are all employed, gives the new employment rate (N +w)/(L + w). This is necessarily greater than N/L as long as N < L. But this also implies that the unemployment rate must fall. And since all OFWs are fully employed, the underemployment rate must  fall as well. Sarah Daway and Jeff Ducanes wrote an article for the Philippine Review of Economics (June 2015) that estimates that the unemployment rate for 2013 would have been 0.4 percentage points lower (i.e., 6.7 % rather than 7.1 %) if overseas workers had been included.

Guess that’s it.

Hope this helps — and that you don’t mind if I put it out on perSE.





(a) This piece was reprinted in Business World 10 August 2015.

(b) For the reply by B. Diokno, follow this link.