One legacy of the 2008 crisis has been that firms have built up big cash piles. Bank of England data show that non-financial firms now have over £425bn of sterling deposits. That's equivalent to over two months of GDP and over 12 months of profits. You might think this is a great comfort, as it suggests that companies can respond to their loss of revenues not just by borrowing but by simply running down these cash piles: isn't that what they are for?

But, but, but. There's a huge problem here. The firms suffering the biggest drops in cashflow might well not be those with the cash mountains. To the extent that the two groups differ, there will be genuine distress.

Which alerts us to a fact overlooked by basic macroeconomics. The economy does not comprise a "representative agent" firm supplying "aggregate demand." Instead, it comprises tens of thousands of different ones, with very different balance sheets (even within the same industry) supplying different goods, who respond differently to changes in economic policy.

This is not a trivial fact. It matters enormously. A given fall in "aggregate demand" which hits cash-strapped firms will be much nastier than one that hits cash-rich ones. We can have no confidence that this crisis will see the latter happen.

Also, as Daron Acemoglu has shown, recessions can be a story of the interconnections (pdf) between firms. The 2008 crisis was so bad in part because banks were key hubs in networks, so when they cut supply everybody suffered.

Networks could amplify this downturn. If a firm fears its customer will go bust, it'll not supply it on credit and so it might well fail: in this way, expectations can be self-fulfilling. Equally, one way in which multiplier effects work is that one firm's failure and one worker's redundancy (or, importantly, one freelancer's lost contract) depresses demand for others.

There's another problem. Let's say you buy a coffee every morning on your way to work, but you then work from home for a month. You won't buy 26 coffees on your day back in the office. But you will have over £50 more to spend on something else. The losers from this downturn will not therefore necessarily be the same as the winners from the upturn. Even if the macro data show a V-shared recovery, the ground truth will be that the upward leg of the V is different from the downward leg.

This wouldn't matter much if capital and labour were fungible - if, say, car manufacturers could easily switch to making ventilators. But they are not. Instead, as Banerjee and Duflo point out in Good Economics for Hard Times, economies are "sticky." Resources cannot switch from one job to another. The pit closures of the 80s and 90s, for example, led to longlasting unemployment and poverty, contrary to the just-so fairy tales of Econ101ers.

Which raises a problem. If unemployed waiters and bar staff cannot easily shift to become (say) the construction workers needed on the government's new infrastructure projects then we might see unemployment co-exist alongside labour shortages.

This is a common effect of recessions. They don't just cut "aggregate demand" but shake up the pattern of it. My chart shows this. It plots the Beveridge curve - the link between unemployment and job vacancies.

As you might imagine, many points lie on a downward-sloping line: in recessions we are at the top-left, with high unemployment and low vacancies, and in good times we're at the bottom right, with low unemployment and lots of vacancies.

But look at the vertical line in the middle of the chart, where vacancies are around 600,000. In 2004, this level of vacancies coexisted with unemployment below 1.5m. By 2008, however, the same level of vacancies existing alongside unemployment of 1.8m, and by 2014 with unemployment of 2.2m. This tells us that the crisis created a greater mismatch between labour demand and supply.

And we've not recovered from it. Those points in the middle of my chart below the main curve all depict the pre-crisis years. Since the crisis, we've moved down the curve, but the curve has stayed further out than it was pre-crisis.

There's a danger that this recession will shift the curve even further out, causing unemployment to coexist alongside vacancies.

Which would mean that the unemployed won't even serve the interests of capital because they'll not bid down wages: unemployed baristas in London would compete against construction workers in the north.

Now, this might not be a severe problem for capital if immigration adjusts, so that we replace one group of migrant workers with another, as baristas return to Poland whilst construction workers come over.

But this is unlikely to be the whole story. Which is why there is an overwhelming case for stopping firms laying off workers in this crisis. The case is not just a humanitarian one, but an economic one; the workers who lose their jobs won't easily find new ones, and won't even be a decent reserve army of labour.