Sunday , December 22 2024
Home / Real-World Economics Review / Growth: weighting the evidence (with changing weights)

Growth: weighting the evidence (with changing weights)

Summary:
In growth accounting, 1+1 might sometimes add up to 1,5 instead of 2. With good reasons. Or at least: with reasons. Let me explain. Economists – and others – tend to look at production when they define  ‘growth’. But we might as well look at consumption. But: what is consumption? It’s the use of stuff. And of services. But economists define it as the purchase of stuff. And services. Which leads to some problems. The value of purchases is, by definition, measured using prices. And prices change – not just in an absolute way but also in a relative way. Stuff that was expensive in 1955, like televisions, might be relatively cheap by now, in 2016. Which means that the purchase of two televisions might add relatively less to total expenditure, measured in fixed prices, than it did in 1955. At least – when these prices are not entirely fixed. Economic statisticians sometimes use 2013 prices to estimate the value of purchases in 2013 and 2014 and to use 2014 prices to estimate consumption in 2014 and 2015 and to link the growth of the ‘volume’ of consumption between 2013 and 2014 tot the growth of the volume between 2014 and 2015. Neat! But this does mean that, over time, goods which have prices which are getting lower in a relative sense are getting lower and lower weights in our estimates of the growth of consumption.

Topics:
Merijn T. Knibbe considers the following as important:

This could be interesting, too:

Merijn T. Knibbe writes Christmas thoughts about counting the dead in zones of armed conflict.

Lars Pålsson Syll writes Mainstream distribution myths

Dean Baker writes Health insurance killing: Economics does have something to say

Lars Pålsson Syll writes Debunking mathematical economics

In growth accounting, 1+1 might sometimes add up to 1,5 instead of 2. With good reasons. Or at least: with reasons. Let me explain.

Growth: weighting the evidence (with changing weights)Economists – and others – tend to look at production when they define  ‘growth’. But we might as well look at consumption. But: what is consumption? It’s the use of stuff. And of services. But economists define it as the purchase of stuff. And services. Which leads to some problems. The value of purchases is, by definition, measured using prices. And prices change – not just in an absolute way but also in a relative way. Stuff that was expensive in 1955, like televisions, might be relatively cheap by now, in 2016. Which means that the purchase of two televisions might add relatively less to total expenditure, measured in fixed prices, than it did in 1955. At least – when these prices are not entirely fixed. Economic statisticians sometimes use 2013 prices to estimate the value of purchases in 2013 and 2014 and to use 2014 prices to estimate consumption in 2014 and 2015 and to link the growth of the ‘volume’ of consumption between 2013 and 2014 tot the growth of the volume between 2014 and 2015. Neat! But this does mean that, over time, goods which have prices which are getting lower in a relative sense are getting lower and lower weights in our estimates of the growth of consumption. This influences our picture of what happened to long-term consumption growth.

Graph 1 shows consumption expenditure by different sectors in France. Together,  government and household and NPISH expenditure  add up to Actual Individual Consumption by households, i.e. consumption of goods purchased by households (like food), goods and services purchased or produced and provided by the government (like primary education and most roads, about 28%) or goods and services purchased and/or produced and provided by Non Profits Institutes Serving Households (NPISH) like churches or your local shanty choir. Graph 1 shows that all categories kept increasing, though there are important differences between periods when it comes to the pace of increase of the different categories. But these patterns are influenced by relative prices as well as by changes in numbers purchased. See also grpah 2. Housing services, part of total services, got much more expensive in almost all countries after about 1965. Which means that housing expenditure gets a larger and larger share in our estimates of the growth of consumption expenditure aside from the fact that we also need more and more house per person (owner occupied houses are the largest exception to the rule that economists do not estimate the use of goods and services but the purchases). We’re supposed to move towards a service economy. Which is true. But our estimates of consumption overstate the case: aside from the fact that we are already amazingly good at producing cheap stuff and are getting better and better at it, services like housing and transport are of course very goods based (messing up with the graphs a bit, as Excel suddenly seems to have a new graph editor).

Growth: weighting the evidence (with changing weights)

By the way – growth (even growth of consumption) is not necessarily an increase of welfare  (though billions do need better health care and food and housing and education). So, let me end with a quote by the Dalai Lama:“We have bigger houses but smaller families; more conveniences, but less time. We have more degrees, but less sense; more knowledge, but less judgment; more experts, but more problems; more medicines, but less healthiness. We’ve been all the way to the moon and back, but have trouble crossing the street to meet the new neighbor. We’ve built more computers to hold more information to produce more copies than ever, but have less communications. We have become long on quantity, but short on quality. These times are times of fast foods; but slow digestion; tall man but short character; steep profits but shallow relationships. It is time when there is much in the window, but nothing in the room.”

Merijn T. Knibbe
Economic historian, statistician, outdoor guide (coastal mudflats), father, teacher, blogger. Likes De Kift and El Greco. Favorite epoch 1890-1930.

Leave a Reply

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