Friday, January 21, 2022

Institute for progress

We need to be reminded occasionally that nothing matters but long-run growth, and long-run growth all comes from productivity, better knowledge of how to better serve human needs and desires.  Yet economic policy is almost all not about growth, but rather about redistribution, and in particular propping up old ways of doing things and the rents associated with them. 

Courtesy Marginal Revolution, the Institute for Progress is a noteworthy new effort to produce growth-oriented policy. Institutions are important, to spread the word and create a constituency for tending the golden goose. 

..productivity growth has been in long-term decline since the 1970s. This is supposed to be the age of ambitious infrastructure investments in the battle to fight climate change, but we can’t even build new solar plants without being vetoed by conservation groups. Hyperloops and supersonic airplanes promise to revolutionize transportation, but building a simple subway extension in NYC costs up to 15 times more per kilometer than it does in other cities around the world. ...The pace of scientific progress has been slowing.... 

(Here and elsewhere see the original for links.) Why? 

Over the last 50 years, we’ve increased the number of veto points at nearly every governmental level, failed to invest in state capacity, and raised the stakes of the debate through polarization. So it perhaps shouldn’t be a surprise that the federal government that went to the moon in 1969 botched production of simple diagnostic tests during a once-in-a-century pandemic.

But the potential is there. Maybe we are not running out of ideas after all, but merely on the edge of technical revolutions, like changing from rail to airplanes:    

...there are genuinely exciting — potentially game-changing — discoveries on the horizon. 

New biotech innovations...may finally deliver cures to some cancers, HIV, malaria, influenza, and EBV. New AI models are unfolding the secrets of the molecular world before our eyes. Electric vehicles are poised for a massive breakout. Commercial rocket launches are becoming such a regular occurrence that they hardly merit news coverage. And record investments are pouring into enhanced geothermal systems, advanced nuclear, and long-duration energy storage, which together could provide the globe with near-limitless clean energy capable of meeting baseload demand.

What to do? 

to do the most good in the world, prioritize issues that are important, neglected, and tractable. When it comes to public policy, that means working on issues that matter a lot, that have only a few individuals or groups working on them, and that are within the realm of political feasibility. Applying this framework to the U.S. context leads to the perhaps surprising conclusion that a lot of the topics that dominate the headlines are not actually the most valuable policies to be working on.

I'll forgive them the electric cars which have no lack of highly visible boosters in the government, and soon, the Fed. But what are the big questions nobody is thinking about? Just asking the question is praiseworthy. Their answers: 

Metascience – How can we change the incentives within science to produce more breakthrough research?

The U.S. federal government is the single largest source of science funding in the world. Despite how much we’re spending, we know surprisingly little about how the structures, incentives, and organizational models within science impact the results we get. But there are signs that something has gone wrong. Principal investigators report spending 44% of their time doing grant-related paperwork and maintenance (as opposed to active research). The average age of a principal investigator working at the National Institutes of Health rose from 39 years old in 1980 to 51 years old in 2008. And our scientific discoveries today appear to be less fundamental than previous advances as we continue to spend more resources. The researcher Pierre Azoulay has argued that we need to “turn the scientific method on ourselves”, and this is the direction we want to push public policy. 

Absolutely! Economists tend to jump to "Science is a public good, subsidize it," but the mechanisms of that subsidy matter. It is possible to pay people to waste time, and for a bureaucracy to become self-rewarding and sclerotic. The science bureaucracy is going woke fast. (The latest: Chemistry journals) At least study what works. Study science funding as we study all bureaucracy and regulation, as a case of public choice theory. Tyler Cowen's fast grants initiative is eye opening as to how much low hanging fruit seems out there.  

Immigration – How can we attract and retain superstar talent from around the globe?

For decades, the U.S.’s asymmetric advantage has been that many of the most ambitious individuals from all around the world desperately want to live and work here...... high-skilled immigration remains a vitally important, yet neglected policy topic. In particular, reforms that leverage the wide scope of executive action are underrated. ... Through both legislation and executive action, the U.S. urgently needs to start taking a proactive approach to recruiting global talent. 

While I absolutely agree on the policy, two warnings. First, counting on "executive action" is dangerous. Parties change every 4 years and the next executive can undo it. Second, knowledge is non-rival. From the point of view of progress, all innovation can happen in China or Europe, and the US can become the following nation. All that matters for growth is that innovation happen somewhere. I'd rather it happen here, and we have for the moment the best place in the world to do it. But this is still about raising the US relative to the rest of the world. 

That immigrants are key to innovation -- even Chinese doctoral students at US universitities, that it is fashionable to disparage, is also not particularly news. But high skill immigration does need a broader political constituency. 

Biosecurity

– How can we prevent future pandemics and accelerate progress in the life sciences?

...the pandemic exposed how weak and ineffectual our public health systems are at controlling highly contagious airborne respiratory diseases. On the other hand, the urgency of the moment — combined with decades of previous investment — enabled us to deliver mRNA vaccines in less than nine months. 

... it’s entirely possible, if not probable, that future pandemics will be even more infectious and deadly than COVID.

In the news today, antibiotic resistant bacteria are on the rise. It might not even be viruses. And I don't know why our enemies are putting so much money into nuclear weapons, when designed viruses could cripple us so cheaply.  

We need to advance metagenomic sequencing capabilities that can quickly detect new viruses. We need to invest in wastewater surveillance systems so we can scale those virus detection capacities across the US. We need to build mRNA vaccine manufacturing capacity so we can rapidly respond when new viruses attack. We need to leverage these systems to fight existing diseases like malaria, HIV, and influenza, which collectively kill millions of people a year. The pandemic has revealed that the life sciences contain much more low-hanging fruit — if we can only implement the right public policies.

Maybe. But we first need to start asking the questions. And then not let it languish, as all the answers to these questions did after the last SARS/ /

There is lots to add. I would add 

Political Economy. How can we reform the US regulatory and administrative state to remove all those veto points, roadblocks, and stifling red tape?  

But I'm here to cheer asking the question, not necessarily the answers that come up in such a preliminary document. 

Perhaps some of this is reflected in 

Policy as roadblock or catalyst

...Regulation can unintentionally block new innovations from reaching the market and rob society of their benefits. Why do we give regulatory special treatment to oil and gas drilling on public land over geothermal drilling? Why hasn’t a single new nuclear power plant opened in the U.S. since the Nuclear Regulatory Commission was created in 1975? Why has the FDA been so slow to approve new therapeutics, tests, and so many other essential goods during the pandemic? In these cases, regulation is the bottleneck holding back private companies from making investments.

They note some successes of federal R&D. The utter failure of NASA's heavy-lift relative to Space-X seems like useful caution. They suggest "pull mechanisms," like prizes. But the deeper question is, why do agencies not use pull mechanisms. 


17 comments:

  1. The share of US GDP for farming is 0.9 percent and all agriculture is ~5 percent. Iowa with the highest share is ~10 percent and more GDP comes from Iowa's research labs than ag. I do an occasional survey with students, colleagues, friends and family and every single one says the share of CA ag gdp is ~ 50 percent and that of Indiana is 60 percent etc. (2.5 & 5 percents respectively). There is no red or blue bias here! All are way off! So during the next primaries, instead of pandering to farmers in Iowa (only 900K independent farmers left in the US and Archer Daniels Midlands type ag companies run 70 percent of farms now), maybe our pols should says how valuable an H1B visa is and how incredibly lucky the US is that it can corner the market of innovative and productive people into its geographic sphere. And just ask any student applying for MIT if they would take up a job or raise a family in Beijing. China or even the US can never replicate the US of the late 20-early 21st century. But the US can maintain the market for innovation and productivity by continuing to do what is does right now i.e. attract global talent.

    ReplyDelete
  2. Sclerosis. Does not look good. Of course the source of the problem is in interest groups and the bureaucracy. Or us! The Reagan revolution has been all but forgotten. Even something as fundamental as the Thatcher revolution in Britain soon petered out.

    A peaceful revolution has the best chance of starting and winning here.

    ReplyDelete
  3. 'Total factor productivity' is not a meaningful metric. If F(K,L) is the production function for the economy, total factor productivity is F(K,L)/(K)/(L) = A(K,L). Total factor productivity growth is dA(K,L)/A(K,L)dt = d(ln(A(K,L))/dt = d(ln(F(K,L)-ln(K)-ln(L))/dt. Suppose F(K,L) = a K^b L^c, then dA(K,L)/A(K,L)dt = da(t)/a(t)dt + (b-1)dK(t)/K(t)dt + (c-1)dL(t)/L(t)dt = g(t), in the most general case. And, K = K1 + K2 + ... + Kn + ..., and, L = L1 + L2 + ... + Ln + ... if K and L are linear combinations of Ki and Li, i = 1, 2, ... , n, ... . But, if capital Ki and Li are not separable, in this sense, then F(K,L) = F1(K1,L1) + F2(K2,L2) + ... etc., and for a society that produces intermediate and primary products in addition to final products, and uses inputs from overseas (imports factor inputs) in addition to domestic inputs, the interrelationships between capital and labor become difficult to untangle. The automobile manufacturing industry is a case in point. It is said that inputs of intermediate products that make up the final product will cross the Canada-US border several times before becoming integrated in the final product in the US. Disentangling the inputs into such intermediate products is difficult, and the "total factor productivity" for such factor inputs is consequently difficult to define. The economist doesn't try to do so--she makes an estimate (a best guess, in most instances) and plugs a number into her spreadsheet model that then crunches the data and presto-magico out pops a number referred to as "TFP". The production environment has become considerably more complex than ever it was in the 1970s. In addition, much of the production is subcontracted overseas while the so-called "high value-added" work is performed domestically (without producing physical product, e.g., "widgets"). Most of the semiconductor chips are designed in "fab-less" design centers while the chips themselves are produced overseas in "fabs" (Taiwan Semi, etc.) The drive is towards zero labor factor inputs onshore, and the transfer of capital intensive production to offshore centers where labor inputs are less expensive. GDP is an imperfect measure in those cases. GNP may capture more of total national output, but imperfectly as well.

    The control engineers would say, with some justification, "If you can't measure it, you can't control it." But, if you're going to measure it, you may as well understand what it is you are measuring because it could be the wrong measure, a.s.

    Suppose one could accurately measure "TFP", precisely. Would that be a sufficient condition, or simply a necessary condition (perhaps one of many necessary conditions). Suppose this new "think tank" staff thought it was a sufficient condition, when it was only a necessary condition, and then went ahead and published its findings and drew up a list of recommended societal changes (actions) based on those (incorrect) findings. What damage or mischief might come of out of that? Enthusiasm is one thing (necessary, perhaps, but never sufficient), and "Web 3.0" is generating enthusiasm in some quarters, as an example, yet I can't help but think that there are too many "big-picture thinkers" these days, (too many chiefs) relative to worker bees.

    ReplyDelete
    Replies
    1. OEE,

      "The control engineers would say, with some justification, If you can't measure it, you can't control it. But, if you're going to measure it, you may as well understand what it is you are measuring because it could be the wrong measure, a.s."

      Absolutely.

      https://en.wikipedia.org/wiki/Total_factor_productivity

      "The word total suggests all inputs have been measured. Official statisticians tend to use the term multifactor productivity (MFP) instead of TFP because some inputs such as energy are usually not included."

      "External costs including attributes of the workforce, public infrastructure such as highways and environmental sustainability costs such as mineral depletion and pollution are not traditionally included."

      "On the basis of dimensional analysis, TFP has been criticized as lacking meaningful units of measurement. The units of the quantities in the Cobb–Douglas equation are:

      A: (widgets * yearα + β – 1)/(caphrα * manhrβ), a balancing quantity, which is TFP

      "In this construction the units of A would not have a simple economic interpretation, and the concept of TFP appear to be a modeling artifact."

      A "better" measure of productivity would simply be Real GDP / Total Credit Market Debt Outstanding. The units are widgets per nominal dollar of debt and both Real GDP and total debt outstanding are measurable quantities.

      Further, regarding your control engineer notation, a government (in particular a Treasury) CAN regulate the amount of debt that it issues irregardless of the wishes of the Legislature.

      Delete
    2. See John's opening statements above:

      "We need to be reminded occasionally that nothing matters but long-run growth, and long-run growth all comes from productivity, better knowledge of how to better serve human needs and desires."

      Not exactly.

      Productivity is a ratio (similar to efficiency in mechanical / electrical systems).

      With either TFP or Real GDP / TCMDO, maximizing productivity does not in and of itself maximize real growth, in the same way that maximizing the mechanical / electrical efficiency of a system does not necessarily maximize the total output of that system.

      Productivity is important because improvements in productivity are non-rivalrous - meaning I don't need to compete with my neighbor for scarce real resources to improve the amount of output that I am generating using what is already available to me.

      See Non-Rivalrous goods:
      https://en.wikipedia.org/wiki/Rivalry_(economics)

      "A good is considered non-rivalrous or non-rival if, for any level of production, the cost of providing it to a marginal (additional) individual is zero. Economist Paul Samuelson made the distinction between private and public goods in 1954 by introducing the concept of nonrival consumption. Economist Richard Musgrave followed on and added rivalry and excludability as criteria for defining consumption goods in 1959 and 1969."

      Country A has an endowment of 1000 tons of iron ore uses that ore to produce 1000 motorized vehicles (1 car per ton).

      Country B has an endowment of 5000 tons of iron ore and uses that ore to produce 3000 equivalent motorized vehicles (0.6 cars per ton).

      Even though country B has produced more cars (more real output), country A has a more productive society (more output for a given level of input).

      Wars are started when countries focus on maximizing output (using goods obtained in conquest) rather than maximizing productivity.

      Delete
    3. "Productivity" = output per unit of factor inputs. If Y equals the output of a Cobb-Douglas production function, F(K,L; A, a, b) = A.K^a.L^b, A>0, a >0, b >0, then, total factor productivity is given by F/(K^a.L^b) = A. Thus, "total factor productivity" equals the constant of proportionality, A. This is a tautology, i.e., "a statement that is true by necessity or by virtue of its logical form" (Oxford Dictionary).

      "Productivity growth" is then (dA(t)/dt)/A(t) and this can be written in terms of factor inputs K and L. Hence, dA(t)/A(t) = dY(t)/Y(t) - a.dK(t)/K(t) - b.dL(t)/L(t), using the chain-rule of differentiation. Productivity growth is non-dimensional.

      Suppose the labor supply is growing at the rate of dL(t)/L(t) = n(t). Further, suppose economic growth is given by g(t) = dY(t)/Y(t). Then,
      productivity growth is given by
      dA(t)/A(t) = g(t) - b.n(t) - a.dK(t)/K(t).

      Suppose dK(t) = -m.K(t).dt + Y(t).dt - C(t).dt, where C(t) is total consumption, all quantities expressed in units of the output commondity for consistency. Then,
      productivity growth is given by
      dA(t)/A(t) = g(t) -b.n(t) - a.m + Y(t)/K(t) - C(t)/K(t).

      Let, y(t) = Y(t)/L(t), c(t) = Y(t)/L(t), and k(t) = K(t)/L(t). Total factor productivity growth can be expressed as follows,

      dA(t)/A(t) = g(t) -b.n(t) -a.m +[y(t) - c(t)].k(t) (1)

      In words, {TFP growth} = {economic growth rate, g} minus [{share of labor, b} x {labor growth rate, n}] minus [{share of capital, a} x {capital depreciation rate, m}] + [{net capital investment per unit of labor, y - c} x {capital:labor ration, k}].

      Any estimation of TFP growth must include the specification of the production function assumed in constructing the regression model.

      Note, in the Cobb-Douglas production function specification, that TFP has units of the output commodity divided by the product of {units of capital}^a and {units of labor}^b, in order to be consistent with the assumed form of the production function. The Wikipedia author's criticism is nonsensical.

      This simple model (incorrectly stated in my original comment posted above) belies the actual complexity of a modern open economy (exports and imports omitted), as noted in that comment.

      Because the definition of total factor productivity is a tautology, it is not uniquely defined, but depends on the mathematical model of the assumed production function.

      Labor productivity, per se, is only valid for a specific level of capital employed. It is a universal metric but its usefulness is limited. Likewise with capital productivity.

      The marginal product of capital is (r + m) on the optimal growth trajectory; likewise, the marginal product of labor is w on the optimal growth trajectory. Off the optimal growth trajectory, marginal products are lower (by definition)

      Determining which of Society A and Society B is "more productive" is more complicated than illustrated in your example. Consider: A BMW Isetta uses less steel per vehicle than does a Suburban, but the Suburban can carry more cargo and passengers per trip-mile than an Isetta can. Which is the "more productive" depends on the nature of the product (passenger trip-miles per unit time per vehicle, for example). It may be "more productive" for Society A and Society B to export their iron ore resources to Society C which has scale economies in the production of "cars", and import such mixes of "cars" as best suits their individual social welfare utility functions. Simple statistics are rarely, if ever, conclusive in determining the winner of "more productive" comparisons.

      Delete
    4. "Determining which of Society A and Society B is more productive is more complicated than illustrated in your example. Consider: A BMW Isetta uses less steel per vehicle than does a Suburban, but the Suburban can carry more cargo and passengers per trip-mile than an Isetta can."

      I presume that in calculations of Real GDP (widgets), the Suburban counts as more widgets than does an Isetta via hedonic adjustment?

      See my orginal postulation - "Country B has an endowment of 5000 tons of iron ore and uses that ore to produce 3000 EQUIVALENT motorized vehicles (0.6 cars per ton)."

      My example was written to illustrate the point out that maximizing real economic growth (Real GDP) and maximizing productivity (ratio of Real GDP and inputs) are not the same thing.

      If Russia does invade the Ukraine and re-absorb it, Russian national Real GDP will likely rise (more workers, more resources, more output). Net Russian productivity may not rise with it.

      Delete
    5. If your example "was written to illustrate the point out that maximizing real economic growth (Real GDP) and maximizing productivity (ratio of Real GDP and inputs) are not the same thing", it didn't achieve that end. To wit,

      Economic growth is dY(t)/Y(t)/dt. Productivity is Y(t)/[K(t)^a L(t)^b]= A(t). The former is the time rate of change of economic activity per unit of economic activity. The latter is the ratio of economic activity to the factor inputs utilized to generate that activity (e.g., your example) at a specific instant in time. The time rate of change of productivity, (i.e., productivity growth) dA(t)/A(t)/dt, is the quantity that is of interest to economists.

      Suppose that in period t = 0, productivity is A(0) and output is Y(0) = F(K(0),L(0)) = A(0).[K(0)^a.L(0)^b], where, a + b = 1 (i.e., constant economies of scale). Now, suppose that you have a brilliant idea on a means to improve productivity in period t = 1, and suppose you are correct and your idea is put to practice resulting in improved productivity, i.e., A(1) > A(0). Suppose that capital, K, and labor, L, in period t =1 is unchanged from period t =0, i.e., K(1) = K(0), and L(1) = L(0). What happens to output in period t = 1? Output should increase, i.e., Y(1) > Y(0) in the ratio A(1)/A(0), ceteris paribus. In this example, productivity growth, ln(A(1)) - ln(A(0)), to a first approximation, leads to output growth, ln(Y(1)) - ln(Y(0)), in equal measure, ceteris paribus.

      Suppose that your goal is not to increase output, but to reduce factor inputs. How would you achieve that purpose? Ans. -- reduce K, and L is proportion to the increase in A. Note that the sum of the exponents is equal to 1, in the assumption at the beginning of this comment. One such evolution, that achieves your goal, is given by Y(1) = Y(0), A(1) > A(0), K(1) < K(0), and L(1) < L(0) such that K(1)/K(0) = [A(0)/A(1)]^(1/a) and L(1)/L(0) = [A(0)/A(1)]^(1/b). You might pursue that goal if your output with the productivity improvement exceeds the demand for your output in the goods or services marketplace in period t = 1, for example.

      Typically, though, the economists prefer to view productivity growth as a driver for societal welfare improvement, especially when net popuation growth is positive, whether by natural increase or a combination of natural increase and increase in net immigration.

      Delete
  4. We need to be reminded that what matters most is equanimity, understanding at the experiential level the transient nature of existence – particles arising and passing away with great rapidity, nothing to cling to, no “I, me, mine,” giving us the ability to face every moment with a peaceful mind.

    ReplyDelete
  5. My other comment not withstanding, I generally agree with your article. About 30 years ago I surveyed a wide range of econometric studies on the optimal level of government for GDP growth: almost all studies found 22% optimal, none was far from that. My experience as a government economic policy adviser in the UK and Australia suggests that governments are increasingly intrusive, and public servants increasingly obstructive. The best governments here - under Hawke (ALP) and Howard (Liberal)- had ministers with wide experience of life, an openness to ideas and a genuine interest in public welfare. Now we have mainly careerists. The situation in many/most Australian universities is also dire, with woke left-wing ideas dominant.

    ReplyDelete
    Replies
    1. Michael,

      "The best governments here - under Hawke (ALP) and Howard (Liberal)- had ministers with wide experience of life, an openness to ideas and a genuine interest in public welfare."

      Is our education system designed to provide that "wide experience of life" or has it become a toxic mix of specialization and career advancement through affiliation (political party, race, sexual orientation, or other)?

      Delete
  6. Current Wealth comes being FIRST in line for the WAVE of CASH being printed!
    Democrats are currently showering THEIR STATES with Trillions.
    It is sad seeing the veritable financial Rape of the next generation. Every Child Born Today has $100,000 of Federal DEBT.
    We are Rapidly copying the centrally planned and centrally controlled economy of the USSR.

    Want to get rich be the son or daughter of Biden and Pelosi et al...where they MAKE millions for the simple action of being the child of an Important Democrat who can arrange meetings.

    Wall Street barely functions to create real jobs...just gamble on High Speed Financial Speculation of Speculation with a FED that offers Unlimited Backing of Speculation for the "right groups"! Or do a renewable tax credit FLIP deal...so a Fortune 500 company can pay ZERO income Tax on multi-billion dollar profits!

    ReplyDelete
  7. John,

    "Political Economy. How can we reform the US regulatory and administrative state to remove all those veto points, roadblocks, and stifling red tape?"

    To reduce the veto points you simply shift to an autocracy (king / queen) instead of the representative Democracy that we live under today.

    That way when the king / queen decides to try, convict, and imprison you for no good reason, there won't be anyone to veto his / her decision.

    ReplyDelete
  8. John,

    "So it perhaps shouldn’t be a surprise that the federal government that went to the moon in 1969 botched production of simple diagnostic tests during a once-in-a-century pandemic."

    The same government went to war in Vietnam in 1965.

    https://en.wikipedia.org/wiki/Gulf_of_Tonkin_incident

    ReplyDelete
  9. https://fred.stlouisfed.org/graph/?g=LbN7
    If you want to improve GDP you will have to improve labor productivity -- we have been in a downward trendline for almost twenty years. We have the most educated laborforce in history, and the availability of labor worldwide; we are not utilizing a resource that is right before us.

    We are more of a service sector economy now so maybe we need to reorganize our business structures -- maybe try going back to hierarchical organizations with an international/national reach -- and also expand the tech capabilities in the service sectors.

    ReplyDelete
    Replies
    1. https://fred.stlouisfed.org/graph/?g=LbN7

      This measure of productivity is a farce.

      If government pays people not to work then hours worked fall and productivity "improves" through fewer hours worked. Inflation rises because income increases while real output stagnates.

      I mean hell, if increasing output PER hour is the secret sauce to improving national productivity, then just pay people to sit at home doing nothing and import everything.

      Hours worked falls to zero and productivity shoots off to infinity.

      Delete
  10. Everyone will be very sad if it turns out that regulations and veto points aren't the actual cause of the problem. Is it true that we have so many more regulations and veto points than Europe that it costs us much more to build infrastructure...than in Europe... by a lot?

    Do we desire to be data driven or are we really happy with preconceptions?

    Maybe we should take a peek at monopoly and monopsony practices, too. We're getting our ass kicked because we let free market practices go too far, not regulation. And, we disdain government but still need to depend on government...what could go wrong there?

    In a country that's exceedingly wealthy, that tolerates legal harassment by the wealthy, which valorizes the individual and disdains the communal... are we really surprised that projects for the common good all grind to a halt?

    Vox had a good article on this recently: https://www.vox.com/22534714/rail-roads-infrastructure-costs-america

    ReplyDelete

Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.