« Infinite equilibrium asset prices? | Main | When should the Bank of Canada raise interest rates? »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Perhaps the question is why do you believe that forecasting models are more accurate than best guesses? Or worth the extra bother?

I think some of the dudes use spreadsheets of multipliers. Do the banks have macroeconometric models? Or should they be using some variant of yield forecasting technology? That will never get them the point forecasts the public demands. Guesses are good.

Didn't McCracken in Ottawa maintain a macroeconometric model back in the day?

In any event, is there a single macro forecaster in the country that does NOT massage the model output? Massaging suggests intuition is at work. Or butt covering.

I have never understood how model-based economic forecasting is possible at all. An economic model typically has endogenous variables Y(t) as a function of contemporaneous exogenous variables X(t), where X(t) includes people's expectations of future variables. But that won't work for genuine forecasts of the future. For that, you need some sort of lags in the model, so that Y(t)=F(X(t-1)), or at least X(t)=G(X(t-1)). But what theory do we have of lags?

In the olden days, under adaptive expectations, an assumed structural relationship between tomorrow's expectations and today's actuals helped economic models forecasts. But that's only because the modeller assumed he could predict what people will expect tomorrow, about the day after, better than those people themselves can.

The are lags in the relation between stocks and flows, of course. Today's net investment determines tomorrow's capital stock. But that only works in discrete time, which fudges the fact that we simply assume that the flow of net investment won't jump in the next second.

Demographics? (Everybody today will be 1 year older next year, if they are still alive.)

Sooner or later someone has to stick a ruler on the time-series data, and just extrapolate by faith and prayer. And the choice of which times series to extrapolate (X or Y), always seems to me to be arbitrary.

This has puzzled me for a long time. I'm glad you raised this question.

I have a question that I wish some enterprising business journalist would ask on a regular basis.

Most business journalists don't have the background to understand why it is important. And I'm not sure as a business reader, I would care. Why should he/I?

Is it so that if some outfit predicts 3% growth +/- 1% and it turns out to be 2, there is some cover?

On political polling, I've seen journalists regularly discussing in minute detail the difference in daily polling of a movement of a few decimal points on a survey that typically has a margin of error of say 3.5% 19 times out of 20. I doubt many are literate on statistics. "Confidence interval" to them means how long the House of Commons is prorogued.

The BBC asked me to come on a radio show and talk about GDP figures the other day (in the morning, so before the numbers were released).

I was wondering if they were going to ask me for a prediction, and what I should say. So I read up on some of the other predictions just in case. I am usually pretty cynical about any kind of 'thinking' work from the private sector, but for some reason it didn't occur to me that I could just make up a number and be just as likely to be accurate as everyone else.

In the end I decided just to talk qualitatively and suggest regression to the mean (as UK GDP forecasts have been especially error-prone recently). Then they cancelled the interview! Still, next time I'll be more suspicious.

I've deleted all comments about income trusts, and will delete others if necessary.

On the origin of species: RRSPs
The avoidance of a condition described by Roland Michener as “invidious”

By: Brent Fullard
February 1, 2010

I think a lesson in history will serve all of us well in evaluating what the income trust issue is about and the great promise that is afforded to all by the arrival of the Marshall Savings Plan, which will see the onslaught of foreign takeover of income trusts halted today, in a way in the which the Liberal Plan is presently unable to, as it would be preposterous to think Harper would embrace the Liberal Plan in Budget 2010 as a solution to the financial pandemic that he has created, in the same way the he would embrace the face saving, yet equally effective Marshall Savings Plan.

Hopefully the Liberals understand the income trust issue sufficiently well to know that it represents the complete defilement of RRSPs as they were first conceived of in 1957......by a Liberal government. Back in those days even the opposition Conservatives spoke in eloquent terms of the need for equality and fairness in providing those Canadians without pensions the same opportunities and level playing field as those with pensions.

I went to considerable trouble to dig out the debate on the second reading of the legislation that ushered in the RRSP in 1957 that was contained in that year’s budget. This material is not available on the internet. Roland Michener called that situation of unfairness as between these two groups of Canadians as an “invidious” one. Invidious meaning:

1. calculated to create ill will or resentment or give offense;
2. offensively or unfairly discriminating; injurious:
3. causing or tending to cause animosity, resentment, or envy: an invidious honor.

Do you suppose that the Conservatives of that day, not to mention the Liberals of that day, under Finance Minister Harris and Prime Minister Louis St. Laurent would find Stephen Harpers’ income splitting for seniors that only applies to those with pension income to be “invidious” or the fact that Flaherty gave the pension funds a bespoke carve-out to placate them such that pension funds can access a companies pre tax earnings and only pay one tax (38%), whereas RRSPs can not, and have to incur two taxes (62%) as being invidious?

I call upon the Liberals and all fair minded Members of Parliament to embrace the Marshall Savings Plan solution, as if it were their own (since there is no pride of authorship) and in the name of these stalwart and morally guided Members of Parliament of the past, and from all sides of the aisle, recorded for all posterity in Hansard below, knowing that The Marshall Savings Plan is a both a litmus test and an IQ test of who today’s Members of Parliament represent: Life insurance annuity salesmen and pension fund annuitants or the 75% of Canadians who simply want equality of treatment

Mr. Michener: “ This [RRSP] could affect some 400,000 taxpayers who are self employed; in fact it could go beyond that as framed because it is open to the employees who are not already contributors to an approved pension fund. I recall that my first plea on behalf of the self-employed described them as being discriminated against by the legislation as it stood in May of 1953 when Mr. Abbott was finance minister. I further believe representations have been made, perhaps over a period of ten years by different associations of self-employed persons, who find themselves in the invidious position of not being able to save for their retirement money which had not been taxed as would have been the case had they been contributing to approved plans.”

Mr. Herridge : “ Mr. Chairman this interesting [RRSP] provision will be to the advantage of possibly, I understand, half a million Canadians, mainly professional and business people. It will, I think, give those people a measure of fair play which had been denied them in the past. We feel that such a scheme is necessary if industrial workers are to have a measure of security on retirement, and the mobility which is their right in a modern democratic civilisation so that they can move from one company to another and not be chained to one area, in other words the right to full security that would be provided by such a national scheme, I think that in view of the large public investment in pension schemes - and a much greater investment still will result from the present changes to the Income Trust Act – the government has a direct responsibility for some action with respect to a national scheme in this country, I recall that last year we were told that over $30 million was lost to the federal government owing to pension schemes, and that sum will probably increase rapidly in the future. Future contributions, deductions and so on will make for even faster growth than in the past and I think that in the present occasion is a splendid opportunity for the minister, in the words of Winston Churchill, to “bring the magic of the law of averages to the rescue of millions who need it at this time.”

Mr. Harris: [on the matter of taxes being deferred under the RRSP legislation for the benefit of all Canadians, rather than simply those with pensions, or what Flaherty has whipped up in 2006 into his tax leakage tirade argument] “On the other hand, there is no question at all in my mind but that stability of income and security for the future do work out to the advantage of the taxpayers of Canada generally, because such security minimizes the number of occasions on which emergency measures might have to be taken here in order to look after persons who otherwise might have been able to look after themselves through this kind of pension program that the hon.member has in mind. I merely say this to indicate that the Minister of Finance is not disinterested about seeing that adequate security and income is provided at all times

The Deputy Chairman: “shall the resolution carry? Some hon. Members: “Carried” The Deputy Chairman: “I think that completes the paragraphs.”

For information on the invidiousnees of the income trust tax, see the comments from Canadians here: http://marshallplan.ca/

Interesting post. I worked at a large metropolitan daily newspaper this past summer and had to do a story whereby I basically called up a bunch of private sector forecasters and asked whether the government's forecasts looked anything like their own. I didn't get into the nitty gritty of the methodology with them, because, quite frankly, the point of the story was whether experts thought the government's revenue forecasts were credible or way out in left field.

From my conversations, however, I got the impression that the big banks did their forecasting in-house with fairly established models. For the smaller outfits, I again was given the impression that people did their own forecasts, but it was a lot more loosey goosey in terms of modelling. One guy basically told me he looked at other people's forecasts and then made a guess.

Interestingly, most forecasters did give bands and were reluctant to commit to point estimates.

Anyhow, I'm not sure where your hunch is coming from, but if you're interested in a reliable answer, why not write a form email asking the basic methodological questions you're interested in, Google the contact info for forecasters you're interested in, and send it off? Might as well go to the source.

But what's their margin of error, and who are these 15?

Top economists see stronger economy

Group of 15 Canadian economists tell Finance Minister they see higher GDP, lower jobless rate than previously expected

http://www.theglobeandmail.com/report-on-business/economy/top-economists-see-stronger-economy/article1453472/

Just visiting: These private sector economists are the first of perhaps several groups that may be consulted for the budget process. Presumably they don't drool.

"They see the economy expanding by 2.6 per cent, faster than their previous forecast of 2.3 per cent." -G&M

Would guess that the margin of error (if calculable) overlaps both forecasts. Trying to second-guess the markets for investment purposes, I pay more attention to how forecasters revise their forecasts than their point forecasts.

Above all, I look at the yield curve. Anyone care to write for a Globe and Mail readership explaining the virtues of yield curve forecasting and restricting forecasts to ordinal outcomes? The catch is that you cannot leave 90% of the readership behind.

In my experience private sector forecasters do their forecasts "judgmentally". That doesn't necessarily mean that there isn't a model floating around there in the background somewhere. Larry Meyer was candid in the 1990s, before he became a governor of the Fed that his forecast was judgmental even though he had the MacroAdvisors model (really, a version of the Fed's old MPS model). Most private sector forecasters thing that forecasting is a marketing tool; they don't take them all that seriously. Indeed Goldman Sachs is very clear that there will be inconsistencies between what their macro guys forecast and what their fixed income people use to make bets in the bond market.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad