From GDP to CIW?

This article below from the Toronto Star looks at efforts to create a Canadian Index of Welfare (CIW). Some folks would like to heave GDP out the window in favour of something that looks like a CIW, but this would be a mistake. A CIW could add to our understanding, but it could not replace GDP. While GDP has many well-known shortcomings in terms of what it does and does not count, it is measured very rigorously. And it is useful when reviewing historical data trends over time, analyzing budgets (and making budget forecasts), and understanding trends in poverty rates.

We still need to decide what would go in a CIW, including how some aspects of it would be measured (environmental imputations, in particular, are challenging). Weights would have to allocated, making the exercise even more subjective. On the latter point, Lars Osberg and Andrew Sharpe created an index of economic well-being a few years ago that allowed the (online) reader to choose the weights based on subjective preference, so someone unconcerned about, say, inequality could set that weight to zero.

I’m on the fence as to the utility of a single alternative number, especially one that would be an index. We would still want to know what part of the index was driving any observed changes. There are also some important questions about interpretation – for example, is a higher divorce rate a negative, symptomatic of “family breakdown”, or is it a postive reflection of the ability of women to leave bad relationships?

While the idea of a single magic number has a certain appeal, this stuff gets very tricky fast, as the delay in producing the CIW shows. Perhaps the real utility of this exercise is to think through what it is that we care about when talking about our well-being. Ultimately we may want a series of sub-indicators across the spectrum rather than one final number.

A richer way of measuring wealth – News – A richer way of measuring wealth

New well-being index would complement traditional GDP

February 19, 2007

Faith and Ethics Reporter
CARDINAL, Ont.–Marj Kempffer hasn’t looked at the sky the same in years. Not since the day nine years ago when the clouds poured heavy sheets of ice on this eastern Ontario village.

Three days it went on, the freezing rain clinging to everything it touched in the worst ice storm to hit North America in decades.

Roofs collapsed, roads became impassable and both electrical wires and the poles that held them aloft snapped. Across eastern Ontario and into Quebec, no one escaped. This village on the St. Lawrence River went eight days without power. Others went for several weeks.

The storm stretched from Montreal to Kingston and caused $1.6 billion in damage. To this day, there are houses that still have not been repaired.

Yet, the Ice Storm of 1998 is the biggest single event in Canadian history to boost the Gross Domestic Product, a simple totalling of all goods and services in the economy that is the most-used measure of the economy.

It’s this irony – that rebuilding in the wake of devastation is good for the economy – that is in large part fuelling an ambitious attempt to produce an alternative to the GDP, one that balances economic growth against a much larger and more comprehensive set of numbers to tell us if we are truly better off. It’s called the Canadian Index of Well-being and its goals are:

  • To reflect a broad range of factors – such as the availability of health care, literacy rates, the quality of air and water, the costs of adequate housing and the value of unpaid work – that together determine the quality of life in Canada.
  • To do it so it’s comprehensive enough to satisfy the statisticians and policymakers but simple enough to be understood by the general public.
  • Give policy makers a tool to show, in quantifiable terms, the positive impact of good social policy such as measures to alleviate poverty, and to demonstrate with hard numbers how a dollar spent now on education or health prevention can reap huge rewards years down the road.
  • “We really need a different kind of statistical indicator – not to replace the GDP, but to complement it,” says Michael Wolfson, assistant chief statistician at Statistics Canada. Neither Wolfson nor any of the other statisticians involved in the project diminish the complexity of the task.

    In health care alone, the CIW’s measure could include ER wait times, rates of cancer and other diseases, body mass index, smoking rates, life expectancy, infant mortality and low birth rates, even rates of depression and suicide.

    Adding all these up, factoring in at least six other “domains” to produce an “integrated index,” would be a monumental achievement. There is still considerable disagreement over methodology (see sidebar) even though the first phase of the project is set for this fall.

    That’s about two years after the original hoped-for launch of the CIW, but organizers say they have preferred to get it right instead of early. The extra time has been used to refine calculations and to co-ordinate the effort with similar projects around the world through the Organization for Economic Co-operation and Development.

    Nonetheless, the CIW has gained a legion of fans, from former Saskatchewan premier Roy Romanow to Bay Street moguls to top government statisticians. At least seven Canadian universities including York and Dalhousie, several federal agencies, and a number of non-governmental groups such as the Atkinson Charitable Foundation are involved in the project.

    To Romanow, the index is an extension of his groundbreaking study of the Canadian health care system almost five years ago, which found that the social determinants of health – poverty, hunger, homelessness, stress and the environment – need to be considered alongside hospital beds and waiting lists.

    For the long-time NDP politician, the GDP’s limitations as a measure of well-being are revealed in the negative inputs it includes.

    Expenditures on cancer treatment, divorce, prisons and funerals are counted alongside factory production and restaurant meals as good for the economy, but few would say such things have improved their lives.

    The ice storm is a case in point.

    A Statistics Canada report months after the storm touted a “rebounding” Canadian economy, based on a surge in the GDP. That surge, it turned out, was thanks in large part to the cost of getting eastern Ontario and western Quebec back on its feet.

    Cardinal Mayor Larry Dishaw shakes his head at the notion that the ice storm was good for anything, much less the economy. The storm devastated local businesses as consumer dollars were diverted to keeping buildings standing and getting the power back on.

    “It was a good six weeks before business was back up,” he says.

    The wait was too long for many. To this day, Cardinal’s main street is dotted by closed stores. Even the local bank eventually closed after 100 years, complaining that business had dropped too much to continue.

    To backers of the plan to find an alternative to the GDP, such stories are all too typical, and reflect how traditional economic thinking ignores the realities that people live out every day.

    “If you don’t measure what counts, what counts is never measured,” says Romanow.

    The effort in Canada is matched by similar projects in countries around the world, particularly in Europe, Australia and New Zealand.

    The statisticians involved in these projects, while they acknowledge the GDP’s faults, say its attraction as a statistic stem from its simplicity. As a totalling of every dollar spent in the economy on goods and services, the GDP can reveal at a glance whether business as a whole is up or down.

    What it can’t tell us is how that money is distributed. A StatsCan report in December found that the gap between rich and poor has been widening in recent decades. Since 1984, the net worth of the top 20 per cent of the population has jumped 64 per cent, while that of those in the bottom 20 per cent has stalled.

    As well, the average net income for the top 20 per cent jumped by more than $20,000 to $110,700 between 1995 and 2004, while it inched up only $400 to $12,200 for the bottom group.

    Over the same nine-year period, however, the total GDP increased by about a third.

    And the more the gap between rich and poor grows, the more inadequate the GDP becomes as a measure of wealth or well-being, says Wolfson.

    “GDP tells us how big the pie is, but not how the pie is divided.”

    Any jump in the GDP that is heralded by economists and policy makers as good news increasingly falls flat with everyone else if they don’t see their own lives getting better, says GDP expert Hans Messinger.

    Messinger made a career of calculating the GDP at StatsCan, and is now spending his retirement working on the CIW as a better measure of growth.

    “That’s what a healthy economy is all about, to not only serve the few,” he says. “The economy is there to serve everybody.”

    Sidebar: How will it work?

    Statisticians, as a rule, like purity in numbers. Much of the attraction of the GDP as a measure of the economy is that it’s a simple measure of one type of number – dollar sales.

    But the sorts of things the Canadian Index of Wellbeing is meant to include are measured in years and percentages, as well as dollars.

    “They’re really apples and oranges,” says Michael Wolfson, assistant chief statistician at Statscan and a top advisor to the CIW project.

    Wolfson has been pushing for the CIW to be expressed as several numbers measuring different things, with the public and Bay Street left to figure out what they mean.

    But former Saskatchewan premier Roy Romanow says the CIW can only really fly if it is easy to understand by Bay Streeters and Main Streeters alike.

    “People should be able to understand at the water cooler what this means,” he says.

    Alex Michalos, social policy researcher at the University of Northern British Columbia and a key member of the CIW group, favours a single number. He cites the “domain” of living standards as an example.

    The first step is to gather eight indicators of living standards – the ratio of top income earners versus bottom income earners, average family income, poverty rate, the economic security index, long-term unemployment, the employment rate, the CIBC index of employment quality and the RBC housing affordability index – for 10 years and plot them on a chart.

    Then he sets the first year for each statistic – 1994 – at 100, a common statistical methodology. The numbers for subsequent years were then recalculated to reflect their relationship to the base year of 1994.

    For instance, family income fell from 47,600 in 1994 to 46,800 in 1995, a drop of 1.7 per cent. That means that if 1994 equals 100 points, the 1995 number is just 98.3. By 2001, incomes had jumped 12 per cent from 1994, so the corresponding number for that year is 112.

    He then did the same thing for each of the other indicators.

    Then, once the numbers are all converted to a common denominator, they could then be averaged to come up with a single number and plotted on a chart alongside the GDP.

    The result is a line on a chart that moves up much more slowly than the GDP over the same period, but which more accurately reflects the state of people’s lives.

    For instance, with the living standard statistics he used, a jump in material wealth as reflected in a 12 per cent increase in family income was tempered by a 10.5 per cent drop in economic security and a minor slip in the overall quality of employment over the same period.


    • While the idea of a single magic number has a certain appeal, this stuff gets very tricky fast, as the delay in producing the CIW shows. Perhaps the real utility of this exercise is to think through what it is that we care about when talking about our well-being. Ultimately we may want a series of sub-indicators across the spectrum rather than one final number.

      That’s my feeling as well. A CIW would have to get around the Arrow Impossibility Theorem before it be meaningful – and that’s not going to happen. But more data can only be a good thing. And I like the idea of a DIY index: it makes it clear just what sort of subjective judgments are needed to construct a CIW.

      (Question: Why not just call it a Social Welfare Function? Isn’t that what they call it in the social choice literature?)

    • Needless to say the index would have many political biases, which I don’t oppose– especially considering our current GDP is latent with those. GDP puts emphasis on growth over well-being. The question is when can we expect Bay St. to care about this (first step: get Parliament Hill to!)

    • If you are going to throw around terms like “Arrow Impossibility Theorem” you could at least perhaps provide a link:'s_impossibility_theorem

      More to the point, I am not sure that GDP gets around the “subjectivity” problem it just imposes one subjectivity as the general objective measure.

      It may be convenient and parsimonious but that does not make it “objective” by any stretch at least by the criteria you all seem to be employing when using the terms subjective and objective.

    • Objectivity implies a detachment of the observer from the object of his or her observation. By definition, it does not exist in economics or other subjects through which people study their own societies.

      Semantics aside, I tend to agree with Marc. The problem is not that GDP is a flawed measure, but that it is often misconstrued as a measure of well-being. GDP is an extremely useful measure of total economic output. The fact that some people misleadingly equate economic output with well-being does not imply a problem with how GDP is constructed. However, work on developing explicit measures of well-being may help to prevent GDP from being misused in this fashion.

    • I would agree with that. But this is not just a question of semantics. Part of the problem is that economists and commentators all too often make the slide from increased output of commodified transactions (GDP) to equating this with an increase in social welfare.

      In this sense GDP is a subjective measure of welfare, but a objective measure of commodified output. Other indexes of welfare could be derived that were equally as objective in that they adequately measure what they purport to measure but are subjective insofar as they are only one of several measures that could be given to measure welfare.

      That said there are a number of externalities generated in the production of commodities which do not get deducted/added to the calculation of GDP which does make GDP only one of several possible measures of “real” output.

      That GDP obtains a large degree of consent among economists and policy makers only means that either Arrow’s math is too tight for the real world or that three good alternative measures of commodified output have yet to achieve assent. Or it could mean that economics is so tightly knit as a profession that it has acted in dictatorial manner and forced its preference on the rest of us: Anyway you slice it, it is still two out three for Arrow.

    Leave a Reply

    Your email address will not be published. Required fields are marked *