Bubble Mongers are more dangerous than bubbles

First, I’d like to start with a quick look at current residential investment in Figure 1. All of these measures are indexed to 1990. The blue line is residential investment/GDP. The other two lines are very broad estimates of the value of residential units under construction (red line) and the value of residential units completed (green line).

2023 is Different

There is a bit of drift here over time. Residential investment (blue) drifts upward in recent decades because in the age of supply deprivation, more of what gets counted as residential investment is brokers commissions (yes, this is counted as part of residential investment) and improvements on existing properties. Units under construction has drifted upward over the past decade because multi-unit projects are taking much longer to complete than they used to.

Figure 1

The main point of interest here is how this cycle is different than previous cycles. Previously all three measures rose and fell together. But, in the current cycle, there has been a significant divergence. This is mainly because residential investment became “overheated” in a way that I don’t think it really ever has before, due to covid supply chain issues. So, the relative value of completions has increased at a relatively moderate pace since 2020. Residential investment originally spiked moderately. But, supply chain bottlenecks meant that all that extra activity got stuck in production queues, waiting for completion. And, so the total value of units under construction ballooned.

Residential investment levelled off, but that was still at a higher rate than units could be completed, so the total value of units under construction continued to rise.

Residential investment has cooled back off, but the rise and fall in residential investment has been almost entirely an interaction with inventories rather than with deliveries and useable homes.

I suppose that means that, in general since 2020, residential investment has boosted GDP more than it has boosted real housing expenditures. And in the most recent quarters, it has been a drag on GDP in a way that doesn’t really reflect lower growth in real housing expenditures.

The drop in residential investment, then, is probably not a reliable leading indicator of cyclical trends. It would have to drop long enough and deep enough to lead to fewer completions, and considering the number of units under construction, the general trend in completions, and the ease with which builders have been selling finished inventory, that seems unlikely to happen in any plausible scenario.

The decline in residential investment along with the seemingly inflated valuation of residential real estate makes this look like the end of a bubble. EHT subscribers know that those are both false signals. But, false as they are, they lead to a certain amount of cheerleading for a “cleansing” recession, and that’s what I’d like to discuss here.

Bubble Mongers are Bad

Here is an excerpt of an interview Milton Friedman gave in 1998.

EPSTEIN: There can be times and conditions in which the stage can be set for malinvestment that leads to recession.

FRIEDMAN: That is a very general statement that has very little content. I think the Austrian business-cycle theory has done the world a great deal of harm. If you go back to the 1930s, which is a key point, here you had the Austrians sitting in London, Hayek and Lionel Robbins, and saying you just have to let the bottom drop out of the world. You’ve just got to let it cure itself. You can’t do anything about it. You will only make it worse. You have Rothbard saying it was a great mistake not to let the whole banking system collapse. I think by encouraging that kind of do-nothing policy both in Britain and in the United States, they did harm.

I will add that the somewhat similar idea of a Minsky Cycle has done a great deal of harm. It is plausible that an economy can get caught up in a froth of over-optimistic investment. And, if it does, obviously some of those investments will produce losses. But, in an economy with central banks run with lots of discretion on unsettled models and federal regulators ruled by the plethora of political biases and imperfections, these cyclical stories with a toe in a plausible real world can be gift wrapped into self-fulfilling prophecies with a confirmation bias bow on the top. Cleansing recessions are the “blood-letting” detritus of 20th century macroeconomics.

The current housing market is a good case in point. If anyone with even a finger on the levers of monetary and political power thinks a recession would be cleansing for American housing markets today, then the body of political economics is simply not developed enough to deserve the power we have let it have over the nominal economy.

The 2008 episode is an interesting puzzle here, because there were a hodge podge of financial currents that all had bubble-like features before 2008. Home prices in some markets that were unsustainably high, some mortgage underwriting that became very sloppy, financial engineering that created products with barely obscured excessive risks, etc. All of those seemingly related financial disequilibria seem like big, fat pieces of a Minsky puzzle that all pointed to the need for a cleansing recession, which is exactly what the country eventually insisted on having. But, as you look more closely at all of those factors, they don’t fit together so well.

One thing that I find odd about cyclical bubble narratives is the focus on debt. Leverage doesn’t seem particularly cyclical or sensitive to interest rates. Figure 2 compares homeowner leverage and corporate leverage over time. To the extent that there are important changes in leverage, they are generally due to sharp changes in equity value rather than pro-cyclical leverage.

Figure 2

But, leaving that aside, what else can we discern about investment and the last few business cycles? Figure 3 compares residential and non-residential investment through several cycles. In the 1970s, high and cyclical inflation at least suggests that there was nominal economic fluctuation that might be described as pro-cyclical, unsustainable inflationary optimism.

Figure 3

The tech boom of the late 1990s and the housing boom and bust of the 2000s seem to have fueled new popularity of Minsky and Austrian business cycle narratives, even though consumer inflation has been famously anchored during this period. Even the recent temporary spike was not associated with unanchored forward inflation expectations.

The Tech Bubble

In terms of residential investment, housing completely sat out the 1990s boom. The nonresidential investment boom and bust was relatively large, and it was associated with a large spike and reversal in the stock market.

In general, a case can be made that the value of a diversified basket of equities in the 25 years since the boom and the rise of various technologies justify the optimism of the time. Anyone who was active at the time knows, however, that valuations were in many cases wildly optimistic. Share prices of some major firms like Cisco Systems and Intel have still not re-attained their peak valuations of the time, and many smaller hardware firms had peak valuations that defied physical production constraints.

This was a bubble. And, it met some, though not all, of the Austrian business cycle characteristics. The boom wasn’t driven by low interest rates or leverage. One of my favorite stories was pets.com, which is one of the iconic failures at the time.

The firm did not go bankrupt, because they hadn’t issued any debt. Petsmart offered to buy the firm, but the offer was less than cash value, so the assets were sold off, and shareholders received 19 cents per share. Optimism. Failed attempts at transformative investments. No debt. There is no straightforward interest rate story to tell here. But there was a sizeable boom and bust - a bubble.

I think three characteristics of this period are worth considering. (1) The decline in tech company valuations led the decline in residential investment. (2) the recession associated with it was relatively mild with trough to peak unemployment of about 2.4%. And (3) investment returned to mid-1990s levels, but not further.

I think there is a bit of siloing that happens on these mis-categorized bubble events. When the tech bubble popped, I think there is an element that we take for granted and then forget about when mislabeling the more recent cycles bubbles. There was no need, and no broad calls for, making up for the bubble investments by engineering an investment collapse. Just because Pets.com didn’t pan out, that didn’t mean that we should reduce investment in vehicle chassis, steel girders, or bagel ovens to make up for it. The loss of Pets.com might have been “cleansing” in some sense, but there was no point in pushing for a more generalized “cleansing”.

The Housing Non-Bubble

I think the fact that the 2000s housing boom definitely was not a bubble led to some destructive inferences because of the way the working definition of a bubble has had to change to fit each new round peg into the Austrian square hole.

First, declining prices didn’t lead declining residential investment. That’s actually the way housing usually goes, which I would suggest is a good reason not to apply bubble models to housing, in general.

The fact that it wasn’t a bubble led to the overestimation of the problem because prices didn’t relent as they would if it was a bubble. A home price bubble didn’t pop in 2006 and 2007 because it wasn’t a bubble. The bust was a choice imposed on the market when nature refused. The bust was self fulfilling prophecy meets confirmation bias. We kept pounding and pounding on the round peg until prices collapsed and it fit into the square hole.

Fortunately, today, (1) so far the Powell-led Fed has been remarkably restrained, and (2) the credit tightening that induced much of the price collapse in 2007 and onward remains in place and can’t induce another trend shift.

Second, bubbles don’t lead to deep recessions. Or, what I frequently see is the explanation that equity bubbles are relatively benign, but housing bubbles are very damaging. First, they affect household net worth more broadly. Second, the use of leveraged mortgages means that the bust infects the financial sector. This gets it half right. Equity bubbles are more benign. But, the real lesson is that bubbles, in general, are benign. Or, maybe more specifically, if we give Volcker his due, bubbles in low-interest rate, low inflation environments are benign.

Forced, pretend bubbles are not benign, for obvious reasons.

Even if we accept the questionable idea that residential investment had been unsustainably high in the 2000s, that rise had reversed before there was ANY recession and before deep price cuts had begun, as is clear in Figure 3. The lesson of 2007 is that a housing “correction”, if that’s what it was, is especially benign. It could have found a bottom before the economy had entered a recession at all!

The lesson of 2008 is that pounding a round peg housing market into a square hole bubble narrative, for years, is extremely not benign.

This is where that siloed intuition of 2000s tech would have been helpful, if applied to 2007 housing. A failed pets.com doesn’t mean that 5 years later you don’t need to invest in ovens or chassis production. In a real investment bubble, this is easily understood.

In the pretend pretend housing bubble this seemed salient. We’d built too much, writ large, because all of the mythological causes of the non-bubble were nationwide - subprime lenders, the GSEs, the Fed, the community reinvestment act, etc.

But, it doesn’t apply to housing any more than to other investments. If you look at the numbers, in city after city after city, before the Great Recession even started, there was an odd decline in residential investment that is idiosyncratic in each city’s time series, but the idiosyncratic negative shock happened in every city at the same time.

Even if you think Phoenix had overbuilt (It hadn’t.), there is simply no reason to expect home production in Milwaukee to collapse because they had overbuilt in Phoenix.

Figure 4 is a random selection of various metro areas with different trend growth rates. You can go to Fred and pick other random cities. They will almost all tell the same story. By 2006, housing production was declining from normal levels. Two years before the recession.

Figure 4

In 2001 and 2002, there was some collateral damage. After the tech bubble bust, investment declined as we worked on getting our cyclical act together. In 2007, it was the equivalent of pushing down the production of chassis and bagel ovens until that pushback finally caused a recession.

A bubble didn’t cause the financial crisis and the Great Recession. Bubble mongers did.

The Covid Cycle

In the current cycle, we did have some inflation. But it’s over. Pretending it’s permanent isn’t going to make the hole round. And there are new disconnects between recent trends and the basic bubble narrative. There has been no general rise in investment. There was a moderate temporary spike in residential investment, which as I outlined above, didn’t actually raise production or production capacity.

If a bubble is the unsustainable investment in new capacity that exceeds potential demand, then this ain’t it. What we have been lacking is capacity.

At the high point of home values in 2022, homeowner leverage was at a 60 year low. Mortgages outstanding compared to personal incomes was at a 20 year low. Debt service as a percentage of personal income was as low as it has ever been measured.

So, yes, in the yool 2023, if you want to pull out the ol’ “recessions are cleansing” narrative, I will be getting all up in your grill, because you are the problem.

Let’s be clear. I am not saying that the point is entirely wrong. There was a recession in 2001. Bubbles might lead to small corrections. Bubbles that are associated with persistently high inflation might require larger corrections. We clearly don’t have that. Expected inflation in 10-year Treasuries topped out at 3% more than a year ago and is below target now.*

I am saying that if this is the way you are addressing monetary and fiscal policy in 2023, after what we have seen over the last 20 years, you are dangerous. You are the problem. And we should hope that a plurality of your fellow travelers do not occupy the seats of power. If there has been malinvestment that needs to be unwound, it has little to do with needing a cleansing recession. We don’t need to stop investing in chassis production in order to make up for spikes in the GameStop share price or nft’s. We certainly don’t need to slow down housing construction.

Put the square hole aside and take your round peg and stick it. . . somewhere else.

* The TIPS spread is currently just over 2%, but TIPS are paid off based on CPI, which tends to run a little higher than PCE inflation, which is the measure officially targeted at 2%.