The Economy Is Booming. Why Don’t Firms Believe It?

This is a guest post on the Odd Lots blog by Alex Williams. Alex is a research analyst at Employ America, and he’s a multi-time former guest on the Odd Lots podcast, once talking about semiconductors and another time talking about municipal finance. He also writes a fantastic newsletter going chapter-by-chapter through Keynes’ General Theory.


Trusting the Boom: Why Aren’t We Seeing Harrodian Instability?

By Alex Williams

One of the cool things about studying the economy is that no matter what’s happening, it’s extremely likely that something similar has happened before. Sure, every time some aspect is different, but by picking out a few types of things to look at – wages, balance sheets, debt levels, gross investment – you can get a little closer to comparing two different species of apple, rather than an apple and a pear.

A corollary to this is that, if you have a Current Event, you can probably find someone in the history of thought who has put together a whole theory around a similar event before. Usually they’ve tried to adduce the particular factors that they feel contributed most, and those can prove a good starting point. The pandemic has proven a great opportunity for this. People have trotted out everything from models of the impact of the stimulus payments on the Treasury General Account to nineteenth century business cycle models derived from situations where pandemics disrupted a harvest.

One of the most interesting aspects of the Covid-19 pandemic has been the success of fiscal policy in preserving consumer spending, and the second-order effects of that success. Rather than an economic crash during the pandemic, we have instead seen aging supply lines straining under a surge in demand, the likes of which hasn’t been seen for decades. Whether in lumber, shipping, semiconductors, or any of a range of industries, this has been a fascinating through-line in recent episodes of Odd Lots.

Separately from my research work at Employ America, I run a Substack whose paid content is a monthly chapter-by-chapter exploration of JM Keynes’ classic General Theory of Employment, Interest and Money. It’s a book that can be difficult to read, even if you work in economics or markets. The language is antique, and the concepts have the diffuse nature characteristic of recent discoveries. However, almost a hundred years on, it is still clearer about most economic dynamics than anything published since. The most recent piece covers chapter eight, about the balance sheet determinants of the propensity to consume. With the above disruptions in mind, I used it as an opportunity to talk through an eighty-year-old paper by Roy Harrod, an economist often forgotten within the simplistic “Harrod-Domar Growth Model” familiar to economics undergrads.

What’s interesting about the dynamical model Harrod presents is that it provides an account of overinvestment during a boom without special reference to financial conditions – an approach that has almost completely exited the public imagination since 2008. However, with consumer demand high, and supply chains struggling to cope, its account of the intensifying feedback loop between consumption and investment should provide a good starting point for understanding the post-pandemic economy.

If you’ll excuse my block-quoting my plain-language explanation of Harrod’s model:

The basic idea is that changes in the level of investment impact the level of present consumption at the same time as changes in the level of investment rely on changes in future consumption for validation. The trouble comes in from trying to relate these two together in a way that makes sense in the long term. This is a really turgid couple of sentences, but what they mean together should become clear soon.

First, imagine there’s a stable growth-investment path. I know that’s a silly idea, but hear me out about how it would work. In the economy, every investment increases productive capacity, otherwise there’d be no reason to invest in it. In a capitalist economy, every investment also needs profits to validate it in the future, usually by having people buy the things that that investment makes. If that doesn’t happen, people give up investing, either because they go bankrupt or it just seems not worth it. At the same time as this is happening, the increase in investment also increases aggregate income, and thus present consumption.

On a balanced path, the increased consumption from the increase in income creates the future consumption needed to validate present investment. Capacity utilization stays stable, while investment, consumption and income all rise stably and in tandem. Everything is perfectly even, everyone gets richer, no one makes a mistake.

The problem is, entrepreneurs use current demand as a big part of forming expectations about future demand. This is natural, and common to Keynes: unless you’ve got good reason to believe something different, then the present provides a pretty good guide to the future.

This is all well and good, but the problem is that the level of investment influences the level of current demand, and the level of current demand influences expected future demand, which in turn influences the level of investment! It’s a loop! If you’re increasing investment in aggregate, that means more folks are being hired on, and aggregate income is rising, so aggregate consumption is rising. As soon as the economy deviates from that stable growth-investment path, the contemporaneous effects of changes in investment on consumption feed back on one another, and start screeching like an amplifier.

So, imagine the level of investment is for one period higher than this stable path. This increase in investment leads to an increase in demand, because the investment creates jobs and income that are spent on consumption. Firms in aggregate will see this unexpected bump in demand, and think “ah! We underinvested in the last period, time to ramp up investment to meet demand!” In the next period, they invest even more, but still see even more new demand, because the added investment has created it. This cycle spirals upward until something bad happens – doesn’t matter what, anything that makes it suddenly really obvious that there’s too much capacity and consumption can’t keep up – and suddenly a bunch of investments aren’t being validated anymore. Everyone sees this and suddenly stops investing, demand craters, and takes employment with it.

Even worse, the same thing happens if you accidentally drop below the balanced growth-investment path as well. If firms invest less than is needed for there to be sufficient future consumption to validate their investment, they will all feel as though they have overinvested in the past, and cut investment further!

In this model, overinvestment feeds overinvestment, and underinvestment feeds underinvestment, and as soon as the system leaves the balanced path, it has no clear way back.

Phew, long block quote. But we should be seeing the effects of this model, right? Consumption numbers are substantially above what most expected this time last year. In fact, some sectors are working far beyond their normal capacity right now. The whole point of a “market economy” for most people is the idea that price signals work by telling folks where to invest for the highest return.

What’s been surprisingly consistent in recent episodes – and commentary from guests afterwards – is that firms don’t trust the boom in demand to last past the transitory disruption. In shipping, the drop-off in demand following the 2008 crash created a vicious economic environment that led to a wave of bankruptcies and consolidation in shipping. Marc Levinson speaks well to this dynamic. For lumber, the problem was in dusting off old mills and making the investments necessary to pull them back online, and then find trucks for the boards.

The reason the dynamic described by Harrod’s model isn’t kicking in yet is that the investment signal from recent demand spikes is not nearly as loud as the signal from over a decade of structurally low demand. While it is hard to know what the future holds, it is easy to expect that it will be like the past, and in the past the demand was almost never there. Sure, our supply chains are rickety and antiquated, often, as Ryan Peterson said, “running off computer systems from the 80s.” However, to update them, firms need to be convinced that demand will be there to validate the investment.

Really what we need is policy so geared towards durably boosting demand (most especially policy that raises wages) that we can move from the last decade of minimal growth to Hirschmanian “unbalanced growth,” and then finally into Harrodian “balanced growth.”

Everyone except 2021-vintage Larry Summers now agrees that the response to 2008 created an environment of low demand that slowed investment, hiring and wage growth for at least a decade. Firms have adapted to this environment, and when faced with a sudden spike in demand, do not remember how to quickly add capacity.

This story is just as bad in the labor market: a decade of asset-light business models, gig labor and low wages have meant minimal investment in productivity enhancement. If wages are low, and there’s not likely to be demand for more products, why bother investing?

While the pandemic was an unlikely catalyst for a demand surge, policy makers should lean into the dynamic and support workers’ demands for higher wages for the demand those higher wages will create. At the same time, the Fed has signaled that it is willing to let average inflation play catch-up while firms re-learn how to invest in capacity. The next step now should be to double down on support for consumer spending and accommodative positioning to prove that the boom is here to stay.

If we can be sure that the final consumer demand will be there to validate new investments, Hirschman tells us, we should welcome the discovery of bottlenecks. They let the economy know where to make investments, and the resolution of one bottleneck leads to the discovery of the next. As long as the Fed is willing to recognize that this process might involve some scattered and transitory price increases, we might be able to see capex really take off. The tricky part now is making sure this process goes on long enough for firms to reset their expectations, and start looking at current demand instead of the memories of a decade of stagnation.

Most importantly, setting things up to deal with climate change is going to involve an unbelievable volume of investment. If we can’t get firms to believe that macroeconomic policy makers will ensure the demand will be there for those investments will be iteratively validated, it is going to be much harder to get enough stuff built.

Everything is a learning experience, and we need to be learning from the fact that firms have yet to start seriously investing when faced with a big jump in capacity utilization. Perhaps the best lesson is that while discovery of these bottlenecks may “cost us” in terms of higher inflation, failing to discover these bottlenecks will cost us far more in terms of lost investment and productivity.

Follow Alex on Twitter and check out his newsletter here.

©2021 Bloomberg L.P.

BQ Install

Bloomberg Quint

Add BloombergQuint App to Home screen.