There’s always a new one lurking. It’s the nature of the beast. Banking is inherently unstable. Most of the time that’s not obvious and banking operations run smoothly. But panics flare up, sometimes for trivial reasons but inevitably a new bank run emerges. Chalk it up to behavioral ticks, irrationality or whatever narrative framing makes you happy. But history is clear: as long as banks have existed, the latent instability roars forth anew every so often.

The latest example is the trio of bank closures in recent days, starting with last week’s voluntary shuttering of Silvergate, followed by Friday’s collapse of Silicon Valley Bank and the decision on Sunday by regulators to shut down Signature Bank.

The question is whether there’s more than isolated turmoil brewing? For the moment, cautious optimism prevails that we’re not looking at a broader banking crisis. But these are still early days and so a full stress test awaits. No one’s really sure how this will play out for a simple reason: the forces that drive banking crises are behavioral, which is to say a sudden spike of fear – fear of losing deposits.

The first priority for policy makers, on the other hand, is crystal clear: lower, and if possible, eliminate the risk of contagion. So far, so good, even if it’s forever controversial. But government intervention at this late date isn’t debatable since history shows what happens if you allow bank-contagion fear to run amuck. The worst-case scenario: the Great Depression. Although there are several factors that led to the worst economic contraction in US history, allowing banks to fail – with no recompense for depositors – was critical in helping turn what might have been a garden-variety if deep recession into a depression.

The hard truth is that banks are different than other industries. Banks are at the heart of economic activity – pumping the necessary liquidity through the system, and it operates on confidence. As a result, when banks fail, the blowback and destruction can quickly spread through the economy like wildfire. Accordingly, banks cannot be allowed to fail in the way that, say, a software firm or a manufacturer of tractors can be permitted to go under.

Why the distinction? Ask yourself a simple question: If you heard that the manufacturer of the car you purchased last year is about to collapse what would you do? Nothing, right? The same question applied to your bank down the street, where you have $10,000 in a checking account, would, presumably, elicit a different response.

Bank runs, as usual, raise politically-charged questions, particularly in the wake of the 2008 financial crisis, when hundreds of banks failed and taxpayer money was used to “bail out” the institutions.

But what do we mean when we say “bail out” the banks? The public’s knee-jerk reaction is to rail against the idea. But the critical lesson in history is that allowing depositors to suffer is playing with fire – with the entire economy at stake.

To be clear, shareholders in banks should suffer, and they are. Depositors are another story.

Critics charge that the government’s decision to ensure that all depositors in the bank failures – even those above the standard $250,000 FDIC limit – is an unfair bailout and that it runs afoul of a free-market economy. Perhaps, but that’s price of banking and preventing a crisis to take down the financial system.

The hard fact is that if you allow depositors to suffer, the risk of contagion goes up. In that case, the potential for a rapid loss of confidence in the financial system writ large arises. At that point, if a crisis emerges, events can easily spin out of control. It’s akin to yelling “fire” in a crowded theater. Nipping that risk in the bud early and effectively is essential to sidestep a wider crisis.

It’s important to recognize that banking, at its core, is an unstable business. Only a small fraction of deposits are sitting in the vault, which creates the potential for trouble if the majority of depositors request refunds all at once – a bank run, as it’s called. Few if any banks can run effectively and remain prepared for such incidents, even though such events arise periodically. That’s one of the rationales for a central bank, but that’s another story.

Banking, in short, is a necessary evil, so to speak, for a free-market economy. Ideally, regulatory policy is optimized so that the risk of bank runs is minimized. By that I mean the incentive to act prudently is maximized.

In the case of Silicon Valley Bank (SVB), poor risk management decisions were clearly a factor. Although SVB held much of its portfolio in safe Treasuries, management favored longer maturities – a losing proposition when interest rates were rising. As a result, the bank was sitting on sizable unrealized losses. That wasn’t an issue, until it was — when all depositors suddenly wanted their money back — immediately.

It didn’t help that the Trump administration loosened the regulatory rules for smaller banks. Did that make a difference? Hard to know for sure, but it was probably a contributing factor that unleashed the current turmoil.

The larger point is that effective regulations for banks is essential, which means that erring on the side of caution is preferable, if only because bank managers will inevitably make mistakes in risk management.

What shouldn’t be tolerated is taking the view that bank depositors should suffer, even if they acted imprudently by concentrating funds in one or several banks. Rather, the time to act on such issues is before a banking crisis, through tighter regulation. Once a crisis emerges, it’s too late to promote the ideals of avoiding moral hazards.

Unfortunately, financial wisdom is unique in the sense that it’s cyclical and not cumulative. Part of the reason banking crises never go away is that the lessons learned in the previous round of trouble are too quickly forgotten. Let’s hope it’s different this time, but I’m not holding my breath.