Jas McKenzie: Empirical Macroeconomist

Jasfest, 7 December 2021

I apologise for not presenting this short tribute to Jas. I have had a tumour removed from my larynx. The prognosis is good, but it will take some time for my voice to recover sufficiently to be able to make presentations.

I want to talk about Jas as a macroeconomist, but we – for I am very conscious of Jas looking over my shoulder – have a wider remit. Good macroeconomics is founded on good empirical analysis. Jas was a good empirical analyst.

It came from the thorough grounding that Jas got at Lincoln under Bryan Philpott; ag economists are generally into reality. When he shifted to Treasury, Jas was put on the land use desk. Alan Stuart, who took over the desk from Jas tells me that he developed detailed data sets. It impressed on Alan the absolute importance of sound data for policy formation.  

Shortly after Jas took over the desk, the New Zealand economy was hit by the 1966 wool price shock. That must have been quite an education. I have read some of the papers written at the time – not the Treasury ones. Economists were flailing around because the single commodity model which economists standard use does not work when there is a major change in the terms of trade – the relative price between products. Jas learnt his trade under fire.

His big opportunity in the Treasury came when Henry Lang asked him to come back from his teaching at Auckland to head what, I understand, was the first serious macroeconomic forecasting group. The story I was told – it was once standard Treasury myth – was that the junior Treasury officials in the macro-team had predicted a huge external deficit for the 1974/5 year, well over 10 percent of GNP, with the implication that Treasury would be involved in a major external borrowing program. As the forecast went up through the layers of the Treasury, each layer cut back the deficit by one fudge or another, each unaware that the layer below had already cut it back. By the time it got to the Treasury Secretary, New Zealand was predicted to run a small external surplus. In fact, the external deficit outturn proved to be well over 10 percent of GDP, just as the juniors had forecast.

We can guess that Henry was furious. He needed an able economist but also one who would not be bullied by those above. He went to Jas. Shrewd move, Henry.

Jas and I really got together when I became director of the NZIER in 1981 and he, now a deputy secretary, was the Treasury representative on my board. I think he liked the forecasting operation at the Institute. I was pouring what resources I could raise into research, which underpinned the forecasting exercise. After I left, I put the research together in Not In Stormy Seas.

Essentially the method, similar to the Treasury one at the time, was that each component of the income and expenditure sides of the national accounts was individually forecast. (We even tried to forecast the production side.) These were not mechanical autoregressive forecasts because the way that they were brought together made them structural ones. The components, although separately forecast, were interrelated and the relationships had to be made consistent.

Using my earlier example, anyone can fiddle the external deficit. But it is not just changing the numbers, say, for exports and imports. Pulling the import number down might impact on stocks, on the composition of consumption and investment and so on; somewhere else in the forecast accounts there would need to be a compensating change.

A major issue is the national savings balance. A reduction in the external deficit means that somewhere else in the economy there has been an increase in savings. Where?

So the macroeconomic team leadership needs to crosscheck these relationships. One becomes aware of what might be called the ‘sociology of forecasting’. The team usually has a slightly optimistic stance and they shift their forecasts in that direction. The crosschecks often expose them. You may find, for instance, that the household saving ratio has jumped to an  historic high, which almost certainly reflects an attempt to restrain the external deficit by implicitly assuming that consumers had voluntarily given up the imports that the forecast had suppressed.

So, leading a forecasting team a strong understanding of empirical realities – for, instance knowing the import composition of gross fixed capital formation which differs, not incidentally, by type – housing, plant, major construction works. It also requires a thorough understanding of macroeconomic theory at a much deeper level than Stage I aggregation and it requires a knowledge of the past. It also involves person skills. Jas had them all.

This is well illustrated by Jas’s involvement on the external Treasury forecasting panel in the 2000s which I was also on. Initially Jas was there as one of the Treasury officials, after having come back to work for them in the infantry – as he put it. When he retired he stayed on as one of the independent panel members.

The external panel arose following an embarrassing budget forecast  in the early 1990s. One purpose was to have an independent assessment of the forecast before publication, another might have been that if they made another embarrassing mistake, the officials could shift the blame a little onto the independent external panel.

The Treasury invited Victoria University to convene the panel. It consisted mainly of academics with macroeconomics – inadvertently demonstrating that there could be an enormous gap between teaching a subject and applying it. At one stage a senior academic had to have explained how GDP was constructed.

Jas really loved the job. He was one of those senior macro-forecasters crosschecking the numbers, but he was also aware that many of the younger Treasury officials had to be taught how to forecast – they certainly had not been taught it at university. I don’t recall him disagreeing with the forecasts so much as nudging them and exploring their implications.

Discussing the Treasury forecasting at that time belongs elsewhere. The forecast of the stock change was primitive, despite Jas reminding us that the inventory cycle was an integral part of the business cycle. We could never do a full savings balance because the panel was never presented with the public sector revenue flows. The forecasting seemed more mechanically autoregressive than in Jas’s day. Jas added a structural insight to the exercise.

Jas’s time on the panel ended when the panel convenor, who knew nothing about economics and even less about forecasting, decided to dispense with the experienced non-academics on the panel. Unfortunately it was just before the Global Financial Crisis so that there was no one left who had much historical memory.  History and structure – that is, theory informed by empirical analysis – need to be heavily drawn upon during a crisis.

In particular the more autoregressive, the more inertia in the forecasts or, as Jas said, the sociology of forecasting finds that forecasters under-predict both major expansions and major contractions. Thus the Treasury forecasts at the time of the GFC underestimated its severity (as did most other forecasts). I am not saying that had Jas been retained on the panel they would have strengthened their forecast contraction, but I am confident that he would have argued that more attention should have been given to the downside prediction (which would have been closer to the actual outcome), which would have sharpened the internal Treasury debate. (Some years later, a Treasury economist told me that had the Treasury had better forecasts at the time of the GFC, they would have given different advice to the incoming Key-English Government, adding that ‘they may not have taken it’. The implication was that had they been, economic management would have been less difficult in the middle years of the 2010s.)

This brief survey has focused on Jas as a macroeconomist. A rounder picture would have given much weight to his interests in international trade theory. However the theory is only marginally involved in empirical analysis (the quants look at future trade scenarios which are rarely subject to the test of comparisons with actual outturns).

Jas had developed his empirical skills in macroeconomic forecasting and (earlier) in agricultural economics. They were integral to his approach as an economist. I leave him with the last word. Whenever we were discussing – not at every lunch – the period which today we describe as the Rogernomics era he would say about the Treasury advocates, ‘I kept asking them for the evidence; they had none.’

Good and Faithful Servant: Jas McKenzie 1939-2020 is an obituary of Jas.