Friday, May 15, 2015

The UK: another case of "productivity problems" where they don't exist

Brad Delong shows us another case where the framing makes us think of the UK economy as low productivity instead of just low NGDP. If we look at it in the same way we looked at the US and Mexico here, then we see there is no actual change in productivity and the Solow production function works quite well ...

The exponents of capital and labor are 1.0 and 0.51 respectively -- comparable to 0.9 and 0.51 for Mexico and 0.85 and 0.40 for the US.


  1. Yes, there is currently a lot of discussion by economists on UK productivity problems. However, the discussion is all very vague.

    Note figure 8 at the following link which breaks down UK productivity growth by sector. It shows that productivity growth varies significantly by sector, so it is worth asking what this means and why the worst performing sectors are worse than the other sectors.

    The worst performing sector is Mining & Quarrying. In the UK that means mostly North Sea oil. This is why I used an oil industry example in discussing productivity in my comment in your previous post. The next two worst sectors are Pharmaceuticals and Finance. Finance productivity reduction is probably down to the reduction in the size of the UK Finance sector over the last few years with the lost jobs being mostly very high paying jobs. However, it’s less clear why Pharmaceuticals productivity is poor. Perhaps it’s related to the cycles of specific drugs? However, that takes us back to what exactly is being measured by macroeconomic productivity. I doubt very much that these industries have forgotten how to do things they could do five years ago.

    Even though many economists are currently discussing UK macroeconomic productivity, I haven’t seen any discussion about sector specifics. I find it disturbing that I can find sector specific data but professional economists discuss only generic economy-wide productivity. I am used to working at the microeconomic level. It is second nature to dig down into data to get more detail. I don’t understand how professional economists can think that they have anything useful to say about productivity if they don’t do this and if it’s not clear what macroeconomic productivity is measuring in any case.

    1. That doesn't make much sense to me, either. Economy-wide productivity is more of an academic concept ... and it's falling in all major developed countries -- the UK isn't special (maybe it's more pronounced there?). It gets the name "the great stagnation" in the US because of Tyler Cowen, but it appears to be just a general property of advanced economies in the information equilibrium model ...

    2. The short list of converging factors resulting in "the great stagnation", "secular stagnation", or "Limits to Growth":

      Long Wave (recurring multi-decade rhythm in the change rate of commodities prices, interest rates, birth rate, and real GDP per capita, driven by demographics, credit/debt, and war);

      Peak Oil (falling net energy per capita from increasing energy cost of energy extraction, resulting in more than 50% of extraction being consumed at unprofitable prices in order to extract increasing supply);

      population overshoot and resource depletion per capita, including water, arable land, forests, and fisheries, reducing the rate of so-called ecosystem services throughput per capita;

      peak Boomer demographic drag effects, causing a once-in-history decline in labor force participation and a shift in the composition of household consumption from high-multiplier housing, autos, and child rearing to low- or no-multiplier spending for property taxes, maintenance, utilities, insurance, and out-of-pocket medical services and medications;

      excessive debt to wages and GDP from "financialization" and deindustrialization of the economy since the 1970s-80s when the US reached peak crude oil production per capita (down 40% since the peak);

      falling rate of profits (a growing share of total profits is from non-productive financial profits and low- or no-productivity, financialized sectors) discouraging productive investment and growth of capital stock in favor of "financial engineering";

      record low for labor's share of GDP reducing the purchasing power of earned income after taxes and debt service, which reduces labor productivity;

      regressive taxation of earned income in exchange for favorable treatment of unearned income from non-productive speculative rentier activities, which discourages productive investment;

      total gov't, private health care and education, and debt service combine for an equivalent of 54% of US GDP, exerting a net cost to the rest of the economy; and

      the convergence of information technology, biosciences, wireless communications, intelligent systems, and Big Data analytics will result in large-scale technological unemployment, disrupting the service sector by accelerating the automation and elimination of millions of jobs without net replacement in retail, hospitality, financial services, health care, and education, sectors in which 65-85% of jobs are held by women. The net loss of employment in these sectors, disproportionately by women, will replicate the loss of job by men in the goods-producing sector since the 1970s-80s.

    3. @Anonymous

      Or the great stagnation is just the increasing likelihood as an economy grows that a given bit of money is facilitating an economic transaction in a low growth market.

      Your short list is a pretty long list! Not sure there is enough data out there to support that complicated of a model ... and even if there was all of those factors are different across developed economies, yet the stagnation appears to be common among them.

      Seems like something more general going on ... Of course, that is just my opinion.

  2. In the Solow growth model, (or any DSGE with capital and depreciation), TFP is usually added to allow RGDP to leave its steady state and grow continually (as well as explain drastic differences in RGDP between countries). The problem that arises without TFP is that, at some point, investment becomes equal to the depreciation of capital every period, so the capital stock and RGDP per unit of labour stop growing. Whether or not their are constant returns to scale is somewhat irrelevant, but the way capital increases isn't.

    1. I completely agree that is the purpose of TFP -- my contention is that those issues are created by the assumptions in the Solow model (real quantities, constant returns to scale). Those Solow model assumptions then create new problems that are fixed with TFP -- that then becomes the bulk of economic growth that has slowed recently. It seems like a game of whack-a-mole where additional pieces of the model solve one problem only to generate another!

      Also, I could be wrong, but in the Solow model I believe there are three sources of growth:

      1. population growth
      2. increased rate of savings
      3. moving towards the equilibrium level of capital (set by savings and depreciation)

      The constant returns to scale limits the exponent of capital to be less than 1, hence allowing for an equilibrium solution in the case where depreciation and savings are fixed rates relative to capital and output, respectively. That is to say constant return to scale is important for the Solow model to have a steady state.

    2. Constant returns to scale are not important for the steady state value of capital, but having the capital exponent be less than one is important. Because of this, your UK model doesn't have a solution, but your US and Mexico models do.

    3. You're right -- in the Solow model where investment is a constant fraction of output and depreciation is a constant fraction of capital. But those are model assumptions ... assuming that the equilibrium is governed by depreciation leads to the requirement that alpha < 1.

      Actually, alpha = 1 would also lead to an equilibrium at any level of K if s = n + g + d (the two lines appear on top of each other).

      However in the model I used to get to the Cobb Douglas function, there is no model of why K is what it is -- it's not really a true "Solow model" and it's fairly agnostic about the processes behind economic growth. Essentially, economic agents perform a random walk and end up in economic states that have higher NGDP because that is more likely than the reverse ...

      In that case, we'd think of K growing because of diffusion at a rate inversely dependent on the size of the state space (slower for larger economies) rather than achieving an equilibrium based on investment and depreciation.

      I think I may have to do a post on that idea ...

  3. Financialization. Once debt and net rentier flows (a kind of "rentier tax") reach a certain level of wages and GDP, the financial sector (and financialized sectors, such as health care and education in the US) becomes a net cost to labor share and profits, and thus to productivity and GDP.

    A secondary effect is extreme wealth and income concentration to the top 0.001-1% to 10% that results in hoarding of overvalued assets at no velocity, coinciding with falling money velocity.

    In the US we are now experiencing a condition in which total annual net flows to the financial sector equal annual GDP output. Moreover, cumulative imputed compounding interest to total credit market debt outstanding at average term is now an equivalent of 100% of GDP in perpetuity.

    After net flows to the financial sector (and financialized sectors), no growth of real GDP per capita is possible.


Comments are welcome. Please see the Moderation and comment policy.

Also, try to avoid the use of dollar signs as they interfere with my setup of mathjax. I left it set up that way because I think this is funny for an economics blog. You can use € or £ instead.

Note: Only a member of this blog may post a comment.