google analytics tracking code

Tuesday, April 30, 2013

Reinhart-Rogoff vs. New Zealand 1951: Waking up from the postmodern nightmare

[Postedit 27 May 2013: For final summing up of these issues, see post Reinhart-Rogoff vs. New Zealand: Final Round?]

[Note: You need to read the two previous posts 1,2, to understand this one!]

The discrepancies between the RR data for New Zealand real GDP growth rates and any published sources have been bugging me since yesterday morning's post. So I retrieved the original RR data from the UMaas site and inserted them into my Excel spreadsheet (yes, I'm also in bed with the devil here).

In turns out I had misread HAP's description of the four missing observations for NZ. They're 1946-1949, not 1947-1950 as I had assumed. RR do not count 1950  in the over 90% category since the NZ debt ratio dipped below 88% that year. If they had, again the result would have been reversed, since that Korean War boom value of +14.67% creeps back in. Thus we have to shift the missing RR data to make a meaningful comparison with the growth values from Maddison/Statistics New Zealand.

I have now done so in this table, where the fourth column now contains the original RR values, 1950 is highlighted in yellow to emphasize that it is not counted, I added 1952 for additional evidence, and the averages are only taken for the relevant range 1946-49 + 1951:

Year Maddison Maddison Geary-Khamis $90 Stat NZ $91/92 RR
1946 8.70 3.37 2.95 7.71
1947 10.91 9.56 0.42 11.93
1948 -9.92 -11.69 3.17 -9.92
1949 10.79 8.54 -5.00 10.79
1950 14.67 12.38 4.97 14.68
1951 -7.63 -9.49 15.60 -7.64
1952 4.34 1.83 -5.50 4.35
average 1946-49+1951 2.57 0.06 3.43 2.57

It's now apparent that the source of RR's growth data is Maddison. (In fact, RR state this explicitly on their data website without providing the data they used directly, so I trust that the UMaas spreadsheet has reproduced this correctly. I'm getting my Maddison values from Statistics New Zealand.) However, even now there are some strange discrepancies. The growth rate for 1946 is 1% too low, that for 1947 1% too high (highlighted in orange). [Postedit as of May 4, 2012: This discrepancy only exists for 1946 & 1947 growth rates calculated from real GDP Maddison data provided by Statistics NZ. The RR values do correspond to the growth rates derived from Maddison's original database. I have no idea why StatisticsNZ provides different "Maddison" values and whether they represent a transcription error on their part or a data revision.] Otherwise, up to slight rounding discrepancies, they are identical. The Stat NZ series looks lagged by one year, but that still does not fully explain the remaining large discrepancies.

Nevertheless, the conclusions of yesterday's post remain. If one takes the Statistics NZ value for 1951 (or include RR's 1950 value), RR's result is stood on its head: highly indebted countries grow faster!

Any result that is so critically sensitive to the inclusion or exclusion of a few observations or a slight time shift between them that decides whether you're in the war boom or the waterfront strike, cannot be taken seriously.

However, at least we can now wake up, if somewhat groggily, from one postmodern data nightmare. Only 999 more to go...


  1. Good work. I found this site on a link from Brad DeLong's blog, and will now bookmark it.

  2. Welcome to the site, JimV. Have you already voted for the Brüning and Mellon Prizes on the sidebar? Polls stay open until May 5 (see the March 27 post). Spread the word.


  3. This may be of interest:

    Note Maddison series is on calendar year basis whereas SNZ series is March year.

    1. Thanks for bringing this FT piece to my attention, it's all grist for the mill (and didn't turn up in my Google searches for background on NZ, although the waterside lockout did, whose effect is still disputed). The calendar year/fiscal year distinction also occurred to me as an explanation but if Statistics New Zealand specifies that somewhere I didn't see it. Mackintosh in his article claims the wool price crashed already in 1951, but my understanding is that the wool boom continued unabated into the 1960s ( The boom was initiated by the US restocking its strategic wool reserves after the Korean War broke out in 1950 (no more cold feet in the Cold War). The 1951 waterside lockout seems more plausible to me, although some people claim its effects on trade were very minor.

      Who would have thought that you would have to become an expert on NZ wool exports and labor relations in 1951 to decide if public debt affects economic growth.

      But now you have to ask, where did Angus Maddison (God bless his soul, he used to have an office just down the hall from me before he died), get calender year data for NZ except from Statistics NZ? And how did he convert their fiscal year data to calendar year?

  4. There is no official GDP data for those dates so these are all unofficial estimates. I think SNZ just reports other unofficial series (eg compiled by Brian Easton). They happen to be on march year basis to be consistent with the later official series.

    Re the wool boom. The volatility matters for the year to year growth rate. The wool price halved in 1951 after a 130% increase the year previous. See here for the data:

  5. Thanks for the wool export price link. If you look closely at their Fig. 1, there is indeed a big spike starting with the Korean War in 1950, but it doesn't decline until 1952, and then to a level twice that of 1949 (where it stays until 1965). It's highest exactly in 1951, when it should have been contributing to an export boom, not a slump.

    These gyrations in the price of NZ's main export would explain the extreme fluctuations in NZ's nominal GDP growth, but not necessarily its real GDP growth if the GDP price deflator reflected wool export prices correctly (which I'm sure it doesn't). And it shouldn't kick in already in 1951 (remember that the Korean War continued for two more cold winters, until 1953). Thus I'll stick with the waterfront lockout as the proximate reason for the 1951 real NZ GDP collapse until a NZ economic historian convinces me otherwise. A volume index would be illuminating, but we would still have to dice up any decline between a drop in real external demand and the waterfront lockout. Looking at the volume of US wool imports might tell us if this is a NZ-specific country effect driven by the lockout or demand.

  6. First, let me say I really enjoyed your three posts on this topic if only because I've spent some time thinking about this issue.

    Now, allow me to share my own thoughts as I hope it may reveal even more of the story:

    I was convinced from the start of the HAP/R&R controversy that the New Zealand part of this story was explained by the 1950-1951 New Zealand Wool Boom and not the 1951 New Zealand Waterfront Dispute.

    The 1952-53 New Zealand Yearbook shows that wool sales were 47.1 million NZ pounds in 1949-50, 107.5 million NZ pounds in 1950-51 and 52.7 million NZ pounds in 1951-52:

    Most of this was caused by a change in price, not a change in output. The average price of wool rose from about 38 NZ pennies a pound in 1949-50 to 88 NZ pennies a pound in 1950-51 and fell back to 40 NZ pennies a pound in 1951-52. (There were 240 pennies to a New Zealand pound.) Production was about 298 million pounds in 1949-50, 294 million pounds in 1950-51 and 315 million pounds in 1951-52.

    According to the HAP/R&R dataset New Zealand's nominal GDP (NGDP) was 1.101 billion NZ pounds in 1949, 1.396 billion NZ pounds in 1950 and 1.446 billion NZ pounds in 1951, so that was a substantial proportion of New Zealand's economy.

    A good paper about the New Zealand Wool Boom is here:

    I can't locate a free copy, nor can I save a PDF file I can cut and paste but I would summarize the episode as follows.

    Demand for wool had been strong since WW II ended but supply had been unresponsive to elevated prices. When the Korean War started in June 25, 1950 there was an immediate elevation in the price of wool. Between that date and March of 1951 the price of wool went up two to three fold depending on grade (lower grades went up more, mainly because that was the kind of wool the military was buying). Demand wasn't simply driven by US military stockpiling as retailers actually used rising prices to induce even higher sales.

    In January 26, 1951 the United States Office of Price Stabilization (OPS) imposed a general price ceiling measure designed to freeze the pre-war price-wage structure. The price ceiling on wool brought trading in Boston (the central US wool market) to a standstill and caused US participation in New Zealand wool auctions to more or less cease. This led to falling New Zealand prices until February 7 when an emergency exemption was granted to the US military through April 1. This caused prices to recover but once the exemption expired prices fell sharply. By June 1951 they had fallen by 50% and by March 1952 they had fallen a total of 70%.

    Now, my sense from reading the history of the Waterfront Dispute is that it was less a strike than a lockout. The government brought in 3000 troops an unknown number of scabs to keep the dockyards running, and thereby crush the union.

    The 1954 New Zealand Official Yearbook shows the Cargo Manifest Tonnage "cleared" (exports) fell from 1,163,934 tons in 1950 to 1,129,629 tons in 1951. It rose up to 1,173,577 tons in 1952:

    In other words the mass of cargo moved fell by only 2.9% and rose by 3.9% the following year.

    What about the actual value of exports? Exports *rose* from 183,752,000 NZ pounds in 1950 to 248,127,000 NZ pounds in 1951, and fell back to 240,561,000 NZ pounds in 1952:

    Wool exports rose from 74,653,000 NZ pounds in 1950 to 128,176,000 NZ pounds in 1951 and fell to 81,998,000 NZ pounds in 1952. Note that wool exports increased by over 70% in 1951 and amounted to nearly 52% of all exports that year.

    So it would appear that the 1951 Dockyard Dispute had little effect on actual exports.

  7. What about the fact that exports increased in the same year that real GDP *fell* 7.6% according to the HAP/R&R dataset?

    This was enormously puzzling to me until I came across your blog posts. Quarterly real GDP for New Zealand that I found seemed to match the HAP/R&R pattern of a decline in real GDP in 1951:

    Note that RGDP declined every quarter of 1951 and did not recover its previous peak until 1954Q2.

    It never occurred to me to check the real GDP data in Statistics New Zealand long term database, although I had done the same for the nominal public debt data.

    As I learned from reading your comments on Matthew C. Klein's Bloomberg article, the Statistics New Zealand real GDP series based on Brian Easton's work is for fiscal years that apparently start in March, whereas Maddison's data is for calendar years and are apparently lagged by a year.

    So my previous research into the 1951 Waterfront Dispute and Wool Boom makes somewhat more sense now. The year 1951 apparently was the boom year, and the recession year is evidently a year later.

    I hope that this is helpful to you as well.

    1. Dear Mark,

      Many thanks for your thoughtful comments. A number of people are now trying to reconcile the Maddison and the StatsNZ real GDP growth rate series. Is the problem based on a fiscal year vs. calendar year discrepancy? If so, did the NZ fiscal year begin in March or July in this period? Carmen Reinhart (, April 27 post) now rejects the Maddison data RR originally used, although the StatsNZ data are even worse for their case.

      Furthermore, we have to be very careful in distinguishing price from volume effects. It's clear that there was a NZ wool export boom starting 1950 and price volatility, but how big were the real vs. the nominal effects and when do they occur?

      As to 1951 export tonnages, I think the waterfront lockout/strike Feb-July (it doesn't matter if you consider it a lockout or a strike) seriously impeded trade, but this might have been compensated for in the second half of 1951 to yield the overall small tonnage decline.

      Another obvious problem is how the statistician converted from nominal to real values. When relative prices are fluctuating so wildly on a short time scale it is well know that this can lead to misleading artefacts in real estimations.

      When I know more about the NZ growth statistics I hope to write another blog that finally clears up this muddle.

    2. A few further points which might be helpful if you continue to investigate the issue. (I am a NZ economist but certainly no expert on the year 1951!)

      1. The "fiscal year" is for the year ended 31 March.

      2. The SNZ website has quarterly wool export price data which shows the price peaked in 1951Q1, and then fell by 60% in 1951Q2. The timing of the wool price drop is therefore coincident with the fall in GDP. I agree with you about needing to be careful about the NGDP/RGDP distinction, although it is still perhaps a somewhat plausible story since (a) terms of trade changes might be expected to induce a volume response (b) nominal shocks matter for cyclical behaviour of RGDP (and this was a fixed exchange rate regime) and (c) the deflators may be mismeasured.

      3. I have no expert knowledge on the economic effect of the waterfront dispute but according to the documentary you linked to I think it was said that the economic cost was NZ£50-£150m. No idea how this figure is derived or whether this was an impact on GDP or something else. But as a ballpark estimate, I note that if NGDP was approximately NZ$1.4 billion, this is equivalent to NZ£700m (NZ changed to decimal currency in 1967 at conversion rate of 1 to 2). Therefore NZ£50-150m represents about 7-20% of NGDP. So it seems this rough estimate is quantitatively consistent with your thesis that the fall in GDP could be caused by the waterfront dispute and ancilliary industrial action.

    3. Coming back to the real vs. nominal distinction in calculating and dating NZ real GDP swings during the wool boom, Mark Sadowski points us to the NZ Statistical Yearbook for 1951-52, which seems to have the only physical volume and price data for this period.

      Here's their table (


      Season. Weight of Greasy Wool Sold. Sale Value. Value Per Pound.
      lb.(000)£(000) d.
      1947–48 280,729 29,375 25.11
      1948–49 287,211 30,882 25.81
      1949–50 297,886 47,138 37.98
      1950–51 293,737 107,509 87.84
      1951–52 314,896 52,734 40.19

      Thus there is a very slight decline in production by weight between the 1949-50 and 1950-51 seasons, but the 1951-52 season has an all-time-high value of 314,896. So in real terms, there is no evidence for a 1951-52 recession. The decline in nominal wool sales is entirely due to the price decline.

      But does anyone know what a 1951-52 sheep season represents, and how it compares with a calendar year or NZ fiscal year (April 1- March 31)?

  8. I think your averages are not accounting for the different timing convention in the NZ Stats data. I think the average would be 2.87 instead of 3.43 with correct timing.

  9. You're right, the StatsNZ data are for fiscal years ending on March 31, while the Maddison data is for calendar years. This was already pointed out by reader Oscar and confirmed by Brian Easton, who first compiled the data (personal communication). But since they are lagging by almost but not exactly a year, there is no obvious way to convert them to calendar year without a month-by-month breakdown, and thereby correct the averages.

    There is a new quarterly real GDP series by Viv B. Hall and C. John McDermott, "Recessions and Recoveries in New Zealand’s Post-World War II Business Cycles", 2011, which offers much better time resolution. However, it is still not clear if the series correctly compensates for the wild swings in wool prices in this period.

    Carmen Reinhart is now claiming on her webpage (, April 27 post) that the Maddison data is incorrect and misdates, when in fact it is the other way around and the Maddison data still seems quite reasonable. It's the StatsNZ data that are misleading if one is not aware of the fact that they are for fiscal years ending March 31.

  10. GS,
    I think it's clear from your analysis that you do need to wake up. Contrary your post, RR do link to the Maddison data (at the U. Groningen source) and it matches the figures that you've attributed to RR (at least at my last download). You've added a comment to my site stating that RR made a transcription error. They clearly didn't and you should correct your post. Here's my full reply.

  11. Dear F.F.,

    The discrepancy is with the StatisticsNZ version of the Maddison data, not the original, as you correctly observe (it makes no difference to the averages, since the +1% for 1946 and -1% for 1947 cancel out).

    I have made a postedit annotation to this post accordingly (see above). So RR are not guilty of a transcription error from the original Maddison data.

    However, why Carmen Reinhart (, April 27 post) now rejects the Maddison data RR used in their 2010 paper is not at all clear. It is the StatsNZ time series that lags by one year due to the use of fiscal years ending March 31, not the Maddison data. The timing of the business cycle in the latter seems correct, although one can question if the price deflators used adequately convert nominal GDP to real GDP in this volatile period.

    In any event, all of the New Zealand brouhaha is irrelevant to the basic question of public debt vs. growth, and just serves to highlight the extreme sensitivity of RR's methodology to historical artifacts.

  12. Anonymous7/5/13 04:25

    Dear Gerald

    Statistics NZ uses, as pointed out, March to March series. There is, obviously, no quarterly series, but

    'A quarterly post-Second World War real GPD Series for New Zealand' (with C John McDermott), New Zealand Economic Papers, 45, 3 (2011), pp. 273-298.

    reconstruct one quite carefully. If you look at the data, you will be able to rebuild a series for calendar years, which from a quick overview, looks close to Madison's.

    Hope this helps

  13. Thanks for the pointer.

    I've already looked at the Hall& McDermott quarerly quarterly series, which shows a similar chronology to Maddison, as you point out. However, because of the better time resolution, we can see that the wool boom actually starts several quarters before the Korean War breaks out in the summer of 1950.

    In any event it is now clear that the Maddison index provided by Statistics NZ has a typo for 1946 (as well as four later years not relevant here), so please use the original Maddison database and not the one available from SNZ until they correct it!