A new idea (really)

I am not convinced by the argument which follows (see Cowen’s Third Law), but I am committed to passing new ideas along, and the researchers — Liu, Mian, and Sufi — have a strong track record.  Here goes:

A unique prediction of the model is that the value of industry leaders increases more than the value of industry followers in response to a decline in the interest rate, and, importantly, the magnitude of the relative increase in value of the leaders versus followers when the interest rate declines is larger at a lower initial level of the interest rate.


The model’s prediction is confirmed in the data.

And finally:

The model provides a unified explanation for why the decline in long-term interest rates has been associated with rising market concentration, reduced dynamism, a widening productivity gap between leaders and followers, and slower productivity growth.

I will ponder this further, anyway here is the link to the paper, 88 pp. long.  If they are right, this is big, big news.  Here is a WSJ summary.

For further background, see also Alex’s earlier post about population/labor force decline and economic stagnation.  It is easier for me to believe that their real interest rate effect is working through the propagation mechanism of population and labor force participation.  Furthermore, I have read too many papers which seem to imply that real interest rates do not much, within normal limits, have a big effect on firm investment decisions.  Their model would seem to imply the opposite, and I would like them to test their implied elasticity against the actual elasticities other researchers have measured.


I suspect weak aggregate demand, which can bring about lower interest rates, is also correlated to less innovation.

Productivity tends to rise when demand is strong and sustained. We may be seeing an uptick right now in the productivity numbers, and we saw a nice boost in the demand-heavy 1990s. The 1960s too.

There is something amiss in the macroeconomics profession.

A good proxy for aggregate demand is nominal GDP. Which countries have experienced the highest growth rates in nominal GDP in recent years? Venezuela and Argentina. How is productivity doing there?

Japan, in contrast, has had some of the highest growth rates in productivity (GDP per working age person) over the last 20 years and their "aggregate demand" growth rate from 1997 to 2017 was zero.

Perhaps I should have said, "real aggregate demand."

Japan, with its heavy QE and zero interest rates and shrinking workforce, is an interesting proposition. It is the only Western nation where young workers are actually starting again to do better in the workforce.

There may be lessons from Japan on immigration and monetary policy. That might be worth a peek from Tyler C.

Very un-PC!


Add on:

China's Labour Productivity improved by 6.85 % YoY in Dec 2017, compared with a growth of 6.49 % in the previous year. China's Labour Productivity Growth data is updated yearly, available from Dec 1953 to Dec 2017, averaging at 7.45 %.

China productivity surging for decades.

I wonder if the main proposition here isn't a variation on what Mike Shedlock has described as an outcome for certain market players having "first access to cheap money", where Liu, Mian, and Sufi are putting forward a model with a coherent mechanism for how that might work in practice. I'd certainly like to see more research down this line to challenge the proposition.


This ties into increased market concentration (in both value and earnings), fewer but larger public companies, increased significance of intangibles in market valuation, and fewer IPOs. Today's start-ups with lots of investment in research and development (as opposed to investment in physical assets) have a disadvantage: all of the investment in research and development that must be deducted from earnings makes it difficult for outsiders (such as lenders) to value. Hence, start-ups have difficulty obtaining funding, either via loans or public offerings, because their valuations can't support either, causing them to be swallowed up by large firms instead. Thus, it's questionable whether a policy of low interest rates actually helps start-ups as opposed to large firms, but policy makers (the Fed) are stuck with low interest rates in order to maintain the confidence of investors and rising (or not falling) asset prices, further increasing market concentration and fewer but larger public firms. Cowen identifies the findings of this study as a "new idea". No, it's not. The new idea would be what to do about the dilemma.

Am I missing something? If you discount the net lead of the leader over the followers at a lower rate, you should obtain a higher value.

A lowering of the long bond rates is associated with an economic contraction. And economic contractions lead to mergers and acquisitions. Follow that path and find that market leaders are in top position during contractions.

GINI, the measure of inequality jumps after each recessions. The rigger for a recessions, not necessarily the cause, is regime change in the presidential election. Follow the Occam's razor, folks. Super wealth gets a new axis of symmetry, their wealth provides marginal liquidity in the primary dealer network for US debt. They charge a fee, we pay the fee.

Cause: any given president dumps a fits load of expenses on the next president who is generally elected to avoid paying those costs. Is in the charts, 90% of our recessions have the same keynesian synchronization. We re told this is a proportional democracy, but it is not, it is a republic; and doing the swindle on that property causes the keynes cycle.

I am skeptical that there is enough data to make a strong statement here. What is the historical period upon which this is based? (paper gated.)

Try this link.


NBER gating for $5 or whatever is infuriating.

Clicking on the link to Tyler's Three Laws from 2015, I was struck at how much better the comments section was.

Tyler, I've seen this happen on many sites: ignoring comment moderation brings down the quality of discussion over time. The crazies gradually crowd out the sensible, smart ones. Have you considered implementing user accounts, an upvote system, moderators, or other measures?

I am not sure an upvote system would work, as the crazies would just upvote each other. Perhaps something like what the NYT does for its opinion pieces would raise the quality of comments, where moderators identify comments as "NYT Pick's", either because they are the best version of a common comment or because they are insightful/informative/interesting?

"Have you considered implementing user accounts, an upvote system, moderators, or other measures?"

I've asked Tyler about that. His reply was that they weren't able to find a simple solution.

Just adding a log in with password would help out. Direct moderation can be very expensive of course.

I'm leery of the NYT's system, because it just creates a bubble via a feedback loop.

Ars Technica had a good idea back in the day (before it became a bubble) of allowing a limited amount of votes. You get 1 up vote a day or 1 per thread or some such. Thus people can't spam vote an entire side of an argument.

All of that requires having some good coder working pro bono or paying someone quite a bit of money.

I would think even a standardized, free-or-cheap, off-the-shelf solution like Disqus (https://en.wikipedia.org/wiki/Disqus) would be an improvement over what we have now. Might be worth a try!

Oh what a tangled web we create
When first we practice to manipulate

It's an oh-duh obvious observation. Nothing new. Institutional inertia (size) coupled with small number threshold (obvious non-logarithmic relative size of numbers) make this something that's obvious and already known to the most casual observer.

I think that increasing concentration has a relationship with decreasing productivity growth through the entrepreneurial channel: higher costs of starting up new businesses (like the world's bank Ease of Doing Business ratings) decrease entry rates and result in less innovation and higher concentration.

My intuition regarding interest rates is that lower interest rates imply in an increased present value of the growth in productivity which implies that returns from present investments in productivity increasing methods are higher. So my intuition is the precise inverse of that paper's results.

Industry leaders should have stronger earnings growth than industry followers. So in the stock market they also haw higher P/Es. Higher PE stocks suffer more from rising rates.

Isn't this all this paper found?

I think that low interest rates, especially in the last 25 years were heavily correlated with a weak economy -- which is arguably just a single observation since we can take the financial crisis as just one macro event. Earlier periods do not represent many multiple independent events also (a point Samuelson made about data for the last century on the stock market).
In weak periods, employers and investors are more risk-averse. This can be seen by looking at the spread in salaries and preference for hiring engineers between the elite programs and ordinary unis. In good periods excellent engineers from anywhere are welcome. In bad periods the value of degrees from Caltech, MIT, Stanford, etc. rise relative to the middle state schools. I don't see why the last recession shouldn't have contributed to the phenomenon. But concentration at the industry level with highly concentrated winners and losers within industries as well as across industries is a much rarer and more recent phenomenon.

Comments for this post are closed