## Misled

Joshua Gans has a theory that I attract more critics than the average academic. Judging by today’s Age, he may be right. Monash University’s David Zyngier is very cross that I was invited to the 2020 summit, and critiques my school productivity study (with Chris Ryan, who for the third time in a row gets off without a critical mention) for being ‘misleading’. This follows on from an article he wrote in the Australian a few weeks ago. I didn’t respond to it at the time, but perhaps now is the point to respond to one issue. Zyngier wrote:

They say government funding for each pupil has increased from \$165 (in today’s dollar terms) in 1964 to \$7169 in 2003.

However, the Australian Bureau of Statistics year book for 1964 indicates that the expenditure for each pupil was pound stg. 193 which, translated into 2003 dollars, is \$3583.

We didn’t use these precise ABS yearbook estimates for our study, since they weren’t directly comparable to later years. But since our nominal dollar spending figures were \$165 for 1963-64 and \$178 for 1964-65, this claim piqued my interest. So I asked my research assistant to go to the library and photocopy the page that Zyngier was referring to. Our best guess as to the page he had in mind appears below.

Â

The 1964-65 figure is \$193.86. One possibility is that this is a coincidence. Another is that Zyngier has confused dollars with pounds.

This entry was posted in Uncategorized. Bookmark the permalink.

### 9 Responses to Misled

1. Andrew – Or is it just the case that you publish far more than other academics and you are media friendly and therefore noticed.

2. David Zyngier says:

Andrew is correct – I actually have confused pounds sterling with dollars for the year 1964. An easy thing to do when checking back over hundreds of pages of data in PDF. Nevertheless, the error remains that Andrew and Chris in their original research mistake \$193 (in 1964 dollars) with \$165 (1964 dollars). When \$193 is converted into 2003 \$ value using the formula recommended by the ABS and available on the University of Melbourne Giblin Library website http://www.lib.unimelb.edu.au/collections/ecocom/giblinfaq.html the 1964 expenditure per pupil is approximately \$1860 in 2003 \$terms buying power – not \$165 in 2003 dollar value as claimed in their original paper.

Moreover the other substantive criticisms of the original paper I have made:
1.

Changes in schooling and the economy
2. Constancy in student demographics
3. More teachers and lower class sizes
4. Data (mis)use

have been ignored entirely by Andrew.

Best regards for your continued research work and participation in the 2020 Summit.

3. Andrew Leigh says:

David, I’m grateful that you’ve acknowledged your error, and look forward to the correction appearing in the Australian.

Of course, I wasn’t implying that this was the only error in your piece – merely the most glaring. You’d know this, naturally, because you sent me a hostile email about the ‘flaws’ in my paper. I wrote back with responses and requests for more detail about where you thought we’d gone wrong. But the next I heard, you’d gone to print in the Oz.

As to the right figure for 1964, I’ve made quite clear that we didn’t use the ABS table above. Instead, we found series that were consistent across the full timespan. But if you insist on using the table above, then I’m afraid I need to point out another error of yours. Our comparison starts with the fiscal year 1963-64. You’ve taken the number for 1964-65. True, \$177>\$165, but now I think you’d have to admit that switching to a different source does not have a major effect on our results.

You’re also wrong to use the CPI. We have a long discussion in our paper about education price deflators. In this case, the bias goes the other way: since education prices have risen faster than general prices, using a CPI deflator overstates the spending increase. We do of course show CPI deflated numbers, but they’re not our preferred estimates.

Similar issues arise in the rest of your comment piece. Here’s a non-comprehensive list:
* We account for many student demographic changes (including NESB) using an Oaxaca decomposition.
* You seem to think that the Oaxaca decomposition is inappropriate here, on the basis that it wasn’t designed for this problem. This misses the point. Oaxaca decompositions have been used in plenty of non-discrimination contexts: they’re just a clear way of separating trends from covariates. A Google Scholar search will show you this.
* Your final paragraph entirely misses the point of the Oaxaca decomposition. We never say that demographics don’t matter. What we say is that the demographic changes since the 1960s/70s have offset one another. Simply put, kids have better-educated parents, and more are from an NESB background.
* True, official data measure student-teacher ratios. But we also provide data on class sizes (footnote 17).

4. Damien Eldridge says:

I have commented on a different aspect of Dr Zyngierâ€™s critique in the following post at my own blog site:

http://economicsgeek.blogspot.com/2008/04/real-versus-nominal-constant-price.html

5. christine says:

Damien: nice post. But Dr Zyngier appears a bit more confused than that even – he cites Andrew and Chris as saying that real expenditures per pupil increased by 243%, but appears to think that this is based on an increase from \$165 to \$7169. Weird.

Personally, as someone who’s had to teach Oaxaca decompositions, I think this is a cool usage. The discrimination thing is a furphy – we are always very careful to say that estimates of ‘discrimination’ based on Oaxaca decompositions are very dependent on what sort of discrimination is occurring (pre market vs post market, and all that stuff).

On attracting critics: Andrew Norton’s probably hit it on the head – if you want Chris Ryan to cop more flak, you should perhaps encourage him to start his own weblog 🙂 And critics are a good thing to have – they can really help improve your work. Of course, that does depend on the quality of the critics.

6. ChrisPer says:

Christine: “that does depend on the quality of the critics.”

(ChrisPer flips evil black cape over shoulder, scowls and slinks into the darkness)

7. John Greenfield says:

Andrew Norton

I rather think that the soft social scientist types have had such a monopoly on public debate in this country that AL’s economic rigour threatens them because they do not have the skills to operate at the same level, Hence all the sneers about all such analyses being “right wing” etc. “We want to live in a society, not an economy” blah, blah, blah.

8. ChrisPer says:

Well, it is HORRIBLY right wing to use actual data instead of PC prejudice.

9. perhaps not says:

Oh dear John Greenfield. Your “analysis” sneers more than those you denigrate. It does you no favours.