'Fraid so, a graph means a wonkish post

That neoliberal economic theory that I ascribe to tells us that economic growth is going to be higher at a time of lower tax rates and greater deregulation of the economy. As those who like high marginal tax rates and regulation never tire of pointing out, economic growth was higher back in those years when we had those things, not neoliberalism. So much for my favoured theory then is one potential answer.

Another is to attempt to work out why this is so. One part of that answer Tyler Cowen seems to think I’m right about. At least, “Tim Worstall gets it” I’ll take as being an indication of that:

There was somewhere between none and fuck all economic growth in the US (and many other economies) in the 1929-1945 period. But the production frontier continued to move outwards, indeed, the 30s are one of the all time great decades for both technology and productivity improvements. The 50s to the 80s were simply playing catch up, in the same way that China and India are now.

There’s more detail in the full post here:

For it’s is saying that the great Post WWII economic expansion was nothing to do with high unionisation rates, Bretton Woods, restrictions upon capital movements, high marginal tax rates, fixed exchange rates or any other of the ”liberal” or ”social democratic” (use one for the US, the second for Europe) theories that are so often advanced.

It was driven by the lack of economic growth in 1929-1945, a lack of economic growth which was accompanied by technological and productivity advances. That is, Post WWII growth happened not because of the policies enacted but despite them.

I would hesitate to insist that this is all of it although I’d give it the old school try if pushed. But it is certainly some of it. No one at all is going to try to say that some of Germany’s growth 1945 to 1955 wasn’t at least in part a recovery from wartime destruction and lows of economic activity. It wouldn’t be difficult to get any economic historian to agree that at least some of the US growth in 1945 to 1955 would have been pent up technological change from 1929 to 1945. Who you’d get agreeing that it was all about that – possibly none. Who you’d get saying none of it was also none.

One of the canonical papers looking at this is from Brad Delong:

This paper addresses the role of macroeconomic policies in determining
long-run rates of productivity growth. We begin by highlighting
aspects of the interspatial and intertemporal variation in
productivity growth which suggest that much of what is important for
raising growth rate lies in the domain of structural policy, since
macroeconomic policies are less than dominant in determining rates
of productivity growth.

Bad macroeconomic policy can prevent or hinder growth, good macroeconomic policy isn’t enough to ensure it. To put this into the crude vernacular that aids my own understanding, not picking your nose on a date might get you a kiss or better but not picking your nose won’t ensure a back seat legover. Or in those more formal terms, a necessary but not sufficient action or the absence of it.

Delong’s paper does point out that slowdown in productivity growth though and that is feeding through into that slower GDP growth.

To switch a little, I wrote elsewhere about inflation estimates:

The problem is that it overstates inflation. The beancounters know this, even if they are loathe to mention it because there is no easy solution to the problem. To understand the issue, look at how we use television prices in our measure of inflation. We now measure different TV sizes. Screens that are larger than 40 inches are included in the basket. This may not seem like an economic issue of great importance, but it is illustrative of why we over-measure inflation.

The biggest price reductions in new technologies — including large televisions — come while they’re still toys for the rich. It’s only when many of us buy something that we include it in the inflation measure – but most of us only buy things once the price has dropped and it is no longer a toy for the rich. When mobile phones first arrived in the UK they cost more than most of the second hand cars I’ve had and airtime alone was pounds per minute. Now they’re more or less an essential. And we’ve all got them because they are cheap. They entered the inflation basket well down the steep cost curve that has taken their price tag from eyewatering to “who cares?”

This problem is not disputed; it just isn’t talked about too much out of embarrassment. US guesstimates are that inflation is overstated by 1 to 2 per cent each and every year because of this.

What follows is speculative, far too much so to ask any editor to actually pay me for writing it down. But I’d argue that this known and agreed over-estimation of inflation is part of our perceived fall in GDP growth and thus productivity.

To understand this you need to grasp how we get to those numbers. We can, and do, measure reasonably well the nominal size of the economy. Near everyone is sending in tax forms, surveys and so on, we can tot up the numbers there and see what’s happening in terms of dollars. But, as we all know – candy does keep getting smaller, doesn’t it? – a $ this year isn’t quite as large as one last year nor last decade. That’s our inflation measure. And so we estimate, as best we can, inflation, to give us real economic growth, the real change in GDP. Productivity is drawn from those changes in real GDP and changes in hours worked. Thus, if we’ve got inflation wrong then we’re getting real GDP growth wrong and so also productivity growth.

No, it’s not quite that simple but that is the essence of it. So, if we’re overstating inflation we’re understating both GDP (real that is) growth and thus productivity growth. We agree that we are overstating inflation so we are understating the other two.

In any one year this doesn’t matter but over decades it becomes important. 1% a year is 100% over 70 years, 2% is a doubling over 35 years. When we’re talking about productivity or GDP changes of 1 and 3 % a year then rolling up such a mismeasurement over a generation makes that hell of a difference.

That doesn’t though explain the difference between the earlier period and the latter. For we’ve nothing to say that the inflation measurement has got worse over that period of time. Except that I’m – being very speculative again – saying that perhaps there is.

To switch again, Paul Krugman:

IT IS A TAUTOLOGY that economic expansion represents the sum of two sources of growth. On one side are increases in “inputs”: growth in employment, in the education level of workers, and in the stock of physical capital (machines, buildings, roads, and so on). On the other side are increases in the output per unit of input; such increases may result from better management or better economic policy, but in the long run are primarily due to increases in knowledge.

Those increases in knowledge are also the same as – technology being the implementation of that knowledge – technological advance. Our inflation mismeasurement comes from the problems we have in dealing with exactly that, the introduction of the new technologies, the application of the knowledge.

Certainly, the western or market economies never did stop adding knowledge nor did technology not advance. But it’s also true that our general and standardised growth accounting would still mark a difference between the two periods, those post-war golden years (up to the mid 70s perhaps) and the more modern neoliberal period. We were educating the workforce much more, progressively more, in that earlier period. We did have women irrupting into the labour force. We were adding more inputs – there are numerous pieces out there showing that Trump can never get to his 3% growth target because we’re not adding workers nor educating them more these days. Demographics are against such high growth rates now, they weren’t post-war.

My argument thus boils down to the post-war era’s growth coming from rather more input increase than technological advance, the neoliberal period’s coming rather more from technological advance than input increases. Yet our known and admitted inflation measurement problem applies not to growth from input increases but only to that from technological advance. I would thus submit that our over estimate of inflation, and our under of GDP and productivity growth, is partly simply because of our known counting problem.

Given that I’ve not the technical skills to even begin quantifying this I’d not strongly argue that this is all of it, all of that perceived slowdown in growth. I would posit that it could be.

It is one of the stylised observations of post-WWII growth, in GDP and productivity, that it was faster in the social democratic days, slower in the neoliberal. We also have this fact of inflation overestimations, that driven by technological advance. Growth in the neoliberal period has been driven more by technology and less by input increases. It seems fair to argue that at least some of the noted slowdown in GDP and productivity growth is due to the known inflation mismeasurement.

How much? How about someone with actual economic skill helping out here?

Support Continental Telegraph Donate


  1. So yoga bottoms have come down in price by about £5 in the last 10 years, have become more stretchy in the right places ( or more comfortable ), and available in a greater range of sizes so more likely to fit. None of this is captured in the index as they have just entered the inflation basket this year. Ok, I think I get that inflation is being underestimated.
    Pork pies and lager in a nightclub are out. It would be interesting to track the prices of everything that has been kicked out over the years, but I digress.
    But did Milton Pbuh say ‘inflation is always and everywhere a monetary event’ or something like that. If true, does it mean that it can be measured without reference to the basket of goods, the ONS bods can all go and work in care homes, and we can measure the GDP deflator without them?
    Not got much help to offer on your question about tech versus productivity I’m afraid.

  2. Milton Friedman called inflation “always and everywhere a monetary phenomenon” and I’ll say it is primarily a monetary phenomenon. But that’s with everything else held constant, as it never is. The Obama years were exclusively borrow-and-spend, resulting in many more dollars sloshing around, but at the same time he used regulation and taxation to torpedo the economy and those dollars moved more slowly than ever.

    Trump’s slashing of the corporate tax rate is massively pro-growth, which is better than a nominal balancing of the government’s books. It will “cost” much less than the official, biased “scoring,” but no one has said it will pay for itself, that it will actually bring in more revenue at its nearly halved rates. Trump’s “infrastructure” slogan, though relying a lot on private business and probably ultimately tamer than his tariffs first seemed, is more government spending. If US producer inflation really is in the mid-2s, interest rates are still held lower, artificially punishing savers, and this must be corrected, no matter how scary Fed rate increases are to the stock market. Tariffs, meanwhile, could torpedo the economy again as badly as Smoot-Hawley did.

    But yes, all our inflation statistics are crap. Basket-of-commodities gauges are crap, as no one eats iron ore, and Supermarket Sweep gauges are too; how much would it have cost 30 years ago to walk around with a communicator and a tricorder in the pockets? Humans have never stopped learning, nor ever stopped applying our knowledge to delivering goods to the customer with less effort, and sometimes obviating the goods and the physical delivery.

    Regarding all the proofs that Trump could not have gotten to 3%, there is proof Trump cannot get over it. Ceteris paribus again. But imagine the blowing up of the Department of Energy, set up 41 years ago to free America from dependence on foreign energy. Its mandate has been overcome by events, achieved despite the DoE. The DoE has no product and does not help anyone else produce anything. Its head, who once vowed to dismantle it, now merely wants to re-bias it back toward oil and gas. Imagine all the productive people whose feet are in bureaucratic concrete being freed to find real jobs. And have whatever monetary policy you please.

  3. PS – The dollar itself used to be our measure of inflation. We had a “basket” with only one thing in it: one-thirty-fifth of an ounce of gold. It wasn’t a particularly good measure, as gold’s fashion and industrial value and its supply varied over the years. But it was immune to tweaking by technocrats with ulterior motives, as none of our current or proposed measures are, until it was overtly abandoned.

  4. “Econometricians are ever so pious,
    But are they doing real science or confirming their bias,
    Their Keynesian models are tidy and neat,
    But that top down approach is a fatal conceit.”

    Fight of the Century:Keynes vs Hayek Round Two–Youtube.

  5. The analysis seems persuasive, can’t help though with the technical quantification though. It is persuasive as I see people definitely much better off than five or ten years ago in the UK although the economic data says not. I maybe see this more than others as I am an expat and infrequent but regular visitor to the UK.

    I assume you saw this though already: https://www.ft.com/content/abc14c66-fb78-11e7-a492-2c9be7f3120a Basically the ONS had held mobile services inflation flat for the last 5 years, due to the fact that they measured the revenue of the firms involved as opposed to measuring the cost of the service provided. If this revision were applied to the UK it would significantly boost GDP growth during the period concerned. One mistake perhaps doesn’t condemn the whole analysis but I does make me wonder about the quality of the rest of the data. This was a particularly prominent part of the economy that everyone knew was changing rapidly, and there is a big on-going debate about what inflation statistics mean, so you would think this area would be regularly checked. But everyone missed it, including those worthies at the Bank of England who are supposed to be controlling inflation. This is meaning we have missed out on some significant growth as a result since if this had been known monetary policy would have been weaker. To say this error has cost the UK economy many billions of pounds is not exaggerating. I guess this eventually comes back to Scott Sumner’s proposal of using NGDP rather than inflation to set monetary policy, with the NGP measure being from market forecasts rather than Central Bank forecasts. NGDP measurements have errors as well but at least then the error would be from wisdom of crowds and there would be a significant price discovery benefit to anyone who discovered a systematic under or over measurement by the masses of this forecast (because they can short the market forecast).