Working with Adam Richards’ METI numbers on the Eikaiwa implosion in Japan

Preface: Japanese government statistics are often filtered in such a way as to make a point or tell a story, so I do take these with a grain of salt. Already, it’s been fairly well established that the Yano Research numbers are a much more accurate portrayal of “all of English study”—including the electronic dictionaries and the like. But let’s use the METI numbers like Mutant Frog blog’s Adam did and see what is going on.

Adam Richards gave a very good blog post, but I just want to add points and present the numbers a little differently. In particular, I am giving the data on a monthly basis because I think it really shows how the business model worked and where the failures were.

On the chart above, I pick out four key series:

1) The blue line is the amount of sales revenue per class, in yen. This is simply the revenue number reported in the month divided by the number of classes held.

Even though classes were usually paid in advance, I like this series because it shows that a lot of revenue usually came in around March, and then there was a steep drop-off. This pattern continues even to recent times.

2) I give the number of classes reported held, by the 100’s. So we are looking at classes held in the hundred thousands over the months. That’s the spring greenish line.

3) I give the number of students reported for the month, also by the 100’s. That is the yellow line.

4) Finally, there is the pink line, representing the number of outlets, which I assume METI is either counting as branches or stand-alone facilities. This is a straight number.

Click to enlarge the graph, by the way.

Noteworthy: The big upset in the data series is when Nova collapsed in 2007. But apparently even before then, there was a shake-up in overall revenue (probably when the government banned it from taking contracts for a period). Nova was going down, but this obviously finished them.

In prior years, there is a surge in revenue per class each spring, followed by a steady rise in number of classes that hangs there for several months–really almost until the next annual cycle is about to start. (The spring green line stays high while the blue line jerks down.)

In the middle of the 2000’s, this pattern starts to break down. The revenue per class still spikes up in springtime, but the number of classes held starts to trend down. This is when I think Nova just started to sell lesson tickets with abandon, even though it couldn’t provide the classes the people wanted. Folks were just being sold empty promises at the start of the fiscal year, so Nova could keep the ruse going.

Once Nova collapsed (the large swings in 2007), you see several changes:

1) The number of students reported studying at Eikaiwas monthly dropped dramatically,

2) The number of classes held dropped similarly,

3) The number of outlets dropped fractionally (even today, it is still higher than it was in the year 2000); and,

4) The revenue per class cost shot up to numbers slightly higher than the peak months under the old (Nova) era.

Adam suggested that this meant paying more for less quality. But I believe quality is not relevant and in fact what it shows is that fewer students have to cover the rent for the industry. As you can see from the trendlines, after Nova, the number of 100’s of classes, 100’s of students, and the outlets all pretty much trend closer together than before (when the students and classes held were substantially over the pink line of number of outlets.)

So this, arguably, was still an industry stretched like a rubber band, even after Nova. Now that GEOS is out, too, it will be interesting to see what the numbers will look like 6 months from now. If charging 13,500 yen a class wasn’t enough to keep the industry solid in the period 2007-2009, what should this style of education really cost?

Can it actually be delivered on a scale business model at all? Which is my suspicion, that it can’t. Otherwise, you’d this this sort of format for education instruction all over the world, and not just in places where the government needs to dump short-term workers, politically, as part of a youth international exchange.

Now, look at this situation: the average class is maybe holding 12 people. You are charging 13,500 yen to hold the class. Not clear how many of the workforce of some 10,000 or so is being employed throughout the week in the whole industry, but it looks like each participant is paying 1125 yen or so a class if it’s 12.

Probably 1/2 the revenue of the class goes to overhead. So 6750 yen remains.

So to make this work, you need 12 people paying regularly, attending each workday, for there to be 33,750 to pay an instructor. Over 52 weeks in the year, that’s 1.75 million yen. But presumably, this is just for one hour’s work a day. And the workday is 8 hours, so an instructor would be making 14 million yen a year.

Not realistic. Not just for the fact that such a number would only employ 2,400 people. So a lot of that workforce has to be part-time. And again, I just take small exception with Adam in saying that the issue is probably not so much about quality—which I have found varies greatly within all “education establishments”—as [it is] to whether there has been enough business in Eikaiwa even to attract talent.

As I pointed out the other day, Yano Research has been letting people know that the whole English instruction industry in Japan is substantially larger than just Eikaiwa. And it looks like, if you tried to filter out the Eikaiwa Implosion, the English teaching business has been stable.

I will blog a bit more about this one later in the week. But if at the height of the last decade, English teaching was an $8 billion business (of which $2 billion was this Eikaiwa), there was a $6 billion industry outside of Eikaiwa. If English teaching has fallen to something more like $7 billion, of which (pre GEOS) $700 million was Eikaiwa, then that leaves a $6.3 billion industry outside of Eikaiwa! (So, one that actually grew.)

If the government tolerates an industry that is simply looking to squeeze it out of you and then send you home, that’s not exactly the story line you want, is it? If so-called businessmen want to cry about how they can’t afford shakai hoken and the like, you want METI’s numbers in the newspaper, not Yano’s, right?

18 thoughts on “Working with Adam Richards’ METI numbers on the Eikaiwa implosion in Japan

  1. “English teaching was an $8 billion business (of which $2 billion was this Eikaiwa), there was a $6 billion industry outside of Eikaiwa.[…]”

    I am sorry, where are those numbers coming from?

    1. They are back-of-the-lunchbag Yano’s from the Japan Times article. The reporting connected with Aeon saying everything is O.K., cites Yano’s FY2005 and FY 2008 figures.

      I want to talk more about this in another post, but I wanted to make the point in this one.

      By the way, your alien registration numbers link was actually for something like arrivals and departures, not resident population. I caught (more than) a little flak for that on Mutant Frog.

      1. My apologies for the previous link. I had no intention of trying to have you caught out and neglected to note it was for only a single month.
        Mea culpa.

        This link should provide overall info, I think:

        08-99-04-0 Excel sheet provides a yearly basis for the numbers in 2008 and it still shows that the number of visa holders in Japan from the US still outnumber the combined numbers for CA/UK/AU combined.

        And I still don’t see where you lunchbag numbers are coming from out of the Yano report. How do you come up with the English teaching being an $8 billion business and Eikaiwa only $2 billion. I can’t find numbers like that in Yano.

        1. Chuckers, I always appreciate your support. I take responsibility for my own relying on the numbers, because they looked very much like ones I thought might be for the Year 2009. So that is 100% me.

  2. Well your numbers are all over the place, I read this about a dozen times before I figured out what numbers you were plugging in where to get what result for teacher’s pay, class costs etc. It might help if you had something closer to the correct figures, so try this (these numbers were typical up to a few years ago at least, I doubt they’ve changed dramatically since):

    Rough average monthly tuition for one student at an Eikaiwa: 10,000 yen, or 2,500 yen per 45-50 minute class, once a week.

    Typical class size: 4-6 students, so 10,000 to 15,000 yen total income for the school per hour.

    Teacher income: typically 250,000 yen per month. The law required (still requires?) a sponsor for a Specialist in Humanities visa to pay 3,000,000 yen per year in pre-tax income.

    Teacher working hours: 25-29.5 per week. 30+ hours per week was avoided by schools like the plague, as then a teacher becomes a true “full time” worker and the school is liable for national health and pension plan payments, plus unemployment insurance. They don’t want to pay, so they keep everyone short. You might get an 8 hour workday for an ALT at a public school, but not for Eikaiwa.

    So figure teacher salary at 2,200 – 2,500 yen per hour. Some schools promise “3,000 yen/hour”, but that is for 20 hour weeks, and they make up the 10,000 yen/month difference in a “contract completion bonus” to get the teacher in at 3,000,000 yen for the year and keep things legal.

    1. You should have to read it about one dozen times. I studied them a good part of the weekend to make heads or tails out of what METI is reporting.

      I am no fan of the PowerPoint world. I think it’s part of the reason that major financial companies implode, and why few people have any sense of up-from-down in many situations. That’s why I put five data series into a four-line graph, and left it monthly.

  3. LB, I am just looking at the figures from METI, courtesy of Adam Richards.

    I know the yen figures and work hours that you are giving are the “conventional wisdom”. But I don’t think the numbers work if you try to build a model using the METI data.

    Plus, there are a lot of unknowns. Rent, for one. I peg overhead at 50%. It might be more.

    Your squib about the 29 1/2 hours is probably right as practice, but visit the General Union’s site for the well-grounded analysis that that’s not even the law. (It is simply an example from a guideline published in 1980.) How do we treat the large number of people who report 40-hour workweeks, even though their contract says 29 1/2?

    You conclude: “So figure teacher salary at 2,200-2,500 yen per hour.” How? We are basing that on the conventional wisdom about the Eikaiwa labor cost structure, aren’t we?

    And how many hours do the people work in the end? What really is the class size?

    My one example, with 12, showed that it can’t be 8-hours of classes with 12. It’s obviously going to be some other combination, and one that leaves the overall industry (at Feb. ’10) unsustainable in the long-run.

  4. The 2,200 to 2,500 yen per hour salary figure is based on a monthly salary of 250,000 yen per month. A quick check of “help wanted” listing confirmed that for a full-time teacher that is about the median going salary now as it was when I taught.

    250,000 yen per month ÷ 4 weeks = 62,500 yen per week.

    62,500 yen per week ÷ 25 hours = 2,500 yen per hour.

    Perhaps things have changed in the last few years, but according to the people I work with who have been working far more recently in the Eikaiwa racket than I have (as in, they’ve been teaching within the last year or two) “25 hours/week” is still the “standard”.

    Which makes sense – you can’t work 40 hours in five workdays when your workday starts (commonly) at 3 or 4 in the afternoon (with kids classes after they get out of school) and end at 9 or, at latest, 10 in the evening (with adults on their way home from work). It is physically impossible to teach 40 hours of Eikaiwa classes with that schedule. Now, the Eikaiwa work longer hours on Saturday, and some offer reduced Sunday hours (mid-morning to early afternoon), but still…

    1. LB, I am in total agreement with you that the numbers you are giving are standard. What I mean by standard is that this is what you see offered, and generally what people think the situation is supposed to be.

      In contrast with that is what I would call the Reality, which is dispatch ALTs working in Japanese high schools for 1,800,000 yen a year and no benefits. Or people stringing four part-time situations together to make 2,300,000 yen. And people fighting over the equivalent of $10/hour gigs to make ends meet. Raiding one-man-shop Eikaiwas to see if they can underbid for any potential private lesson students (like the infamous Steel Frame housing of the Philippines king, Sterling Bryson, who supplemented his vast third-world construction fortune with Eikaiwa gigs (2004-??)).

      The reality of Eikaiwa numbers, and the false promises and fraudulent contracts put out there for the Immigration Department are, I would say, two different things. This is why I am studying the METI numbers.

    1. Actually, the thing you might want to be concerned about are the posts that don’t seem to be telling me very much. The only comments I have from you are these short ones.

  5. My apologies, folks. Still mastering my iPhone.

    My point again, and in brief this time, is this analysis based on “filtered” data – and aren’t they often so (pace Disraeli) – is a useful insight, but hardly, I suspect, the whole dismal/positive picture depending on ones perspective in relation to the picture: Highly qualified and experienced instructors on the slide to MacDonald’s style pay and conditions. And, “mega”-employers like Benesse Holdings via their subsidiary Berlitz seeing, in the former’s case, record profits whilst subsidiaries like Phoenix Associates record 20% current downturn on profits and almost certain further instructor pay-cuts (am not sure yet about the state of play at Simul).

    Add into the mix, SKYPE type lessons of uncertain quality; back-packer private lessons; increased dispatch employed instructor input into universities at a fraction of the standard rates and with additional, unpaid duties expected (eg, grading, test creation and assorted admin and additional reports), then the overall picturej of the detestable, in my opinion, “eikaiwa” “industry” grows even more, counter-intuitively, profitable, at least in the short-term, for the Benesse/Berlitz whales (despite strikes and lawsuits in Japan and customer discontent in Germany) whilst the “benign” and “innovative” smaller organizations attempt to remain in play, and communicative competency as measured against global criteria, eg the Common European Framework (CEF) and ALTE (Association of Language Teachers in Europe) as well as TOEFL, continue to plummet in contrast to China, S Korea and India.

    Nice money if you can get and accrue it for pedaling a pernicious illusion to, perhaps, a willfully gullible public fixated stubbornly and, even now, expensively on a game of charades.

    Recommendation: Sell standard eikaiwa. Buy whales if cheap, but monitor and be ready to dump rapidly. Truly long-term, some genuine innovative stuff starting to come down the pipe on a par with global “platinum” standards (product/process-wise): Observe and invest where they operate, in domestic Japanese terms but to rigorous global standards, out of the box if you can still afford to.

  6. Short and to the point. Is, sometimes, to the point.

    Lengthy and uber empirical is lengthy and
    – often – outdatedly faux statistical: Data dead and rotting and uber faux analytical.

    Current pay scales seem, and are currently, ranging from ¥900 – ¥1500 ph for contact hours (quality of instructor input disregarded and admin unpaid).

    Some organisations above this payscale giving warning that paycuts, one guesses, to compete with competitors’s competitiveness soon to be a done deal. Like it or leave.

    Statistical analysis vital and useful, but, often outdated and trees-focussed – where they, previously, stood – and not on the changing forest’s shape.

    Prediction using a Kenneth Galbraith literary economics approach: The faux industry of “bei-eikaiwa” and, subsequent, “eikaiwa” is dead at all levels (university pretensions included). And the limited statistical mapping simply, Sci-fi like or shamanistically, rattles the bones after the fact. With “sexy” physicalia intruded. Extruded.

    Update your field-research data.”
    Some current pay-rates: ¥900 – ¥1500 ph

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s