Today the United States sits in the midst of the largest wealth bubble in post-World War II history, as measured by household net worth (or wealth) relative to gross domestic product. As I showed in detail recently in the Journal of Business Economics, only two other postwar bubbles come close, with peaks in 1999 and 2006, just prior to the tech stock crash and the Great Recession.

No one should ignore the risk that this bubble will burst, as did the previous two, with declines in the value of stocks DJIA, +1.03% SPX, +0.66%  , homes, and other assets accompanied by recession, unemployment, and disruption in the plans and lives of many Americans.

We’re not off the hook, however, should the bubble fail to burst or should peak valuations decline only gradually.

Household wealth is now more than 5 times the size of gross domestic product, the largest wealth bubble yet.

Either way, rates of returns on assets likely will remain significantly lower than in the past, creating its own set of risks for pension plans, personal retirement planning, new homeowners, commercial real estate, and other investments, especially those dependent upon returns extrapolated naively from the recent past.

If valuations stay at current high levels, the expected return is closer to today’s 0% to 2% real interest rate on bills and bonds or a 5% earnings rate on corporate stock, but not the 7% or 10% total returns on a diversified portfolio many people have become used to receiving.

$20 trillion bonus for investors

In addition, even without a sudden crash, there likely would be a gradual reversal of the extraordinary period of “bonus appreciation” that, since the mid-1990s, entailed gains of more than $20 trillion in household net worth over and above what might have been expected had wealth merely increased at the same rate as income.

Suppose that the wealth-to-income ratio simply returns to some prior average. Combined with the lack of past appreciation, that could a represent a negative swing of $40 trillion or more (again, at today’s level of GDP) relative to projections that recent past growth is prologue.

The recent three bubbles share characteristics and causes unique to a modern period since about 1990.

Many conventional theses associate the 1999 bubble with tech stocks and the 2006 bubble with owner-occupied housing. True, those markets were at the forefront of asset appreciation just prior to the two recent recessions with which they are associated, but, more generally, appreciation stretched across most asset markets during those times.

By analogy, consider a water table beneath a foundation that continually rises to new heights and recedes only to new higher lows. Previously the water found weaknesses in one set of cracks (tech stocks), then another (housing). Same cause, but a different point of initial break-through before the flooding affected the entire structure.

With the water table higher than ever, how sanguine should we be that the old break-through points have been sealed and that no new ones will be found?

Conventional wisdom wrong?

The unique quality of this modern period also warns us that conventional wisdom about our inability to time the market may be wrong. That wisdom relies upon long periods of investment, as when stock investment prior to the Great Crash of 1929 produced a high return over a long period of time. Or research that indicates that next year’s return is not easily predicted by last year’s.

But it doesn’t tell us what to do when we have a unique situation. We’ve had only two other sets of data points in which we reached a peak near to this one. And they both involved a crash. Those investors who came out best in those prior cases were those who did diversify more into safer assets during the peak appreciation period.

The evidence. At the end of the third quarter of 2018, based on the Financial Accounts of the United States as estimated by the Federal Reserve, the ratio of net worth to GDP reached a level of 5.3. Just before the Tech Bubble Recession, the ratio hit a record 4.5 in the first quarter of 2000. Before the Great Recession, the ratio hit a record 4.9 at the end of the first quarter of 2007.

In both prior bubbles, the crashes led to a drop in the value of net worth to about 4 times GDP. Even that level remained high relative to prior history, since in no single quarter before 1998 had the household net worth-to-GDP ratio ever reached 4.0 or higher.

Let’s suppose that the current bubble pops back to that same high trough of 4.0. That means a drop of about one quarter in the value of all household net worth. Drops to past historical averages would mean an even larger hit.

By limiting downturns in wealth valuations to new and higher plateaus, monetary and fiscal authorities may have made investors more secure and added to risks in new ways.

Such possibilities fit neatly into Merton and Bodie’s thesis that “well-intentioned government policies aimed at reducing the systemic risks of a crisis in the global financial system may have the unintended and perverse consequence of actually increasing the risk of such a crisis.” Or Hyman Minsky’s analysis of why stability is destabilizing.

Does that mean that past interventions were wrong? By no means. But it does mean that they have entailed new sets of risks — in this case, a higher level of wealth valuation from which a fall can be even greater.

Harder to diversify

The modern period differs not only in the size of total wealth bubbling but its extension across asset markets. As a result, the recessions that occurred in this modern period were preceded by uniquely high amounts of bubbling. By way of contrast, much of the 1970s saw stock declines as real estate rose (relative to GDP), and the 1980s saw the reverse.

This means it’s harder today to diversify against a downturn that can hit almost all markets.

So, something different clearly is going on in this modern period with its unique serial bubbles.

Actually, a lot of things, but with a common characteristic: a lot money has been sloughing around and into the United States — sources ranging from unprecedented rises in government debt-to-GDP and Fed forays into purchasing mortgage-backed securities to sovereign banks, investors, and oligarchs abroad attracted by the stability of the dollar and a rule of law.

Add in rising profit levels and, not unrelated, the increasing sophistication of planners to arbitrage financial markets and the tax system (declaring costs but not gains) when after-tax, after-inflation rates of interest run close to zero or negative. And, since the Great Recession, the dollar becoming the best-looking horse DXY, -0.09%  in the shoe factory.

But can it continue? I suspect that the best of macro policy makers will soon discover that this modern, roller-coaster version of a Tower of Babel cannot reach ever heavenward, though I suspect they will try as they try to limit the size of new downturns and crashes in wealth.

So here is the best advice I can give.

For federal policy makers to be able to engage in future stimulus policy, especially in any future recession, they must gain credibility by engaging in a budget policy that schedules a reduction in debt-to-GDP over time, such as was in place in the period after World War II.

Active investors and lenders need to understand that tomorrow’s successful investments will differ from the past, and highly leveraged investment will not be bailed out by unusually high levels of appreciation for good and bad alike.

For passive investors, I can’t tell you what are the best investments, but certainly keep aside enough safe and relatively risk-free investments both as protection against a crash and as a source of investment funds later.

Eugene Steuerle is a fellow and the Richard B. Fisher chair at the Urban Institute. He is co-founder of the Tax Policy Center of the Urban Institute and Brookings Institution.