Follow Andrew on Twitter.
In the main political discussions about economic questions this week, two important related fallacies have been repeated. Here I seek to unpack and expose them.
Here's the first: "If the government taxes the rich to give money to the poor, that increases GDP."
A number of commentators took exception to my article about austerity earlier this week, in particular my claim that there is no robust economic theory according to which fiscal transfers increase GDP. Daniel Knowles of the Economist stated of me: "he alleges that transfers can't increase AD, which is totally wrong" and that "In theory, taxing people with a low marginal propensity to consume & giving more to those with a high MPC" increases GDP. (The "marginal propensity to consume" or MPC is the percentage of income that we consume as opposed to save.) This is a remarkably common thought, based on a confusion in the way that simple macroeconomics is often taught. It features in many political discussion, also, where it is suggested that cutting benefits "sucks demand out of the economy".
The naive idea behind this is the intuition that if people spend more of their money, rather than saving it, that will boost growth. Except that's just wrong. In orthodox macroeconomic models, money that isn't consumed is invested. And investment boosts GDP as well as consumption. So when the economy is operating properly, even if we could costlessly transfer money from high savers to high consumers and would not distort their incentives by doing so, there would be no effect upon GDP – it's just that a bit more of that GDP would be consumption and a bit less investment.
To get anywhere, therefore, this idea has to posit that the economy is subject to "over-saving", in the sense that some of what is saved is simply set aside for speculative or liquidity purposes and not actually invested into anything. We also have to posit that that extra ability to speculate and that extra liquidity isn't itself going to increase investment or consumption – the money is just (from the social point of view) pointlessly being not used (even if not using it might in some obscure way benefit the individual).
Well, spelling out how any of that could really work and why is pretty complicated and highly disputed, but there are indeed models (in particular, Keynesian models) in which it is assumed to occur. So let's park our objections and concede for now that there can be over-saving. Does that mean we can increase GDP by switching money from those that save a lot to those that consume a lot?
No! For at least three reasons. First, it isn't the amount of saving that counts here, but the amount of over-saving. The fact that one person saves a lot doesn't mean that she over-saves – it might be optimal for her to save a lot. And the fact that someone else hardly saves at all doesn't mean he doesn't over-save – ideally he might save even less! So transferring money from low MPC folk (say, "the rich") to high MPC folk (say, "the poor") might increase over-saving, rather than reducing it.
To make the concept of transfers from over-savers to non-over-savers work we'd need some systematic way to identify those that over-save and give money to those that do not over-save. Simply stating "I'll take it from an over-saver and give it to an optimal saver" isn't describing a policy – just an act. To make it a policy, we'd have to set out how we find these inefficient over-savers. And the government simply lacks any way to do that. It has no idea what the optimal level of saving even is, let alone who is saving at that optimal level and who is not.
The second problem is that even if it could find over-savers and efficient consumers, it cannot guarantee how they will react to funds being transferred between them. For example, an over-saver subject to a extra tax might react by saying "Golly! Things are even worse than I thought and the risk of unexpected demands on me even greater! I'd best save even more." Conversely, the person that saves very little might react to being given a transfer by saying "Finally! A bit of cash! But who knows when I might get more? I'll set this aside for a rainy day." In that case the transfer results in even more over-saving, because both the taxed and the beneficiary react to the transfer by increasing their over-saving.
Third, even if we could find over-savers and perfect savers, and even if the way they reacted to the transfer would not offset its effect, the GDP increase would still have to exceed the highly non-trivial costs of the process. The deadweight cost of taxation is estimated by the Treasury at around 30% of the value of the tax.
Transfers are done for what economists call "equity" reasons – they are there to help the poor or the sick and so on. That's good, but the process of doing it intrinsically distorts incentives, which if anything must reduce GDP. Whether you believe in a high benefits or low benefits society, it's just daft to believe that raising benefits will make the economy grow faster.
Here's the second fallacy: "When calculating the growth in GDP, we should net off the growth in debt."
That fallacy featured, for example, in the latest The Deep End's "Heresy of the Week" and is quite common in right-wing discussions about the state of the economy. The intuition behind the thought here goes rather like this. If you see me consuming more one month, does that mean I've got richer? If my consuming more is because my income has risen, then all well and good. But if I'm consuming more because I've taken out a large loan and am blowing it on short-term consumption, then I have really become richer. Again, suppose I buy a house – am I now richer? No, because I've had to part with money to buy the house. But suppose I didn't use my own money? Then we must net off any debts I took on to buy the house in assessing my ultimate net worth.
So if, say, GDP rises £1 but debts rise £5.40, the thought goes, we aren't really £1 better off at all, because we must later save up something extra – foregoing some consumption – to pay off the £5.40. Are we not, in fact, £4.40 worse off?
Well, no. That's just wrong. Let's see why. Suppose that in country ABC land there are three persons, A, B and C, who all work and produce £1000 of output. Then the GDP of ABC land is £3000. But suppose we discover that that A has borrowed £100 from B, B borrowed £100 from C, and C borrowed £100 from A. Then it turns out that as well as £3000 of output, ABC land has £300 of debt. Does that make us think the "true" GDP or wealth of ABC land is any less? No, because the debts of the ABC dwellers are all to each other. The collective debt of ABC to anyone else is zero.
Hence the relevant debt-related question about the wealth of a country is not how much debt its citizens owe to each other, but, rather, the net debts its citizens owe to or are owed by the rest of the world.
The fact that Britons increased their debts to each other during the 2000s may well be a reflection of their unrealistic expectations of the future, and might increase internal stresses in the financial system and damage the future growth in the economy. But it does not make the GDP accumulation at the time any less real. That's just wrong.
Follow Andrew on Twitter.
In the main political discussions about economic questions this week, two important related fallacies have been repeated. Here I seek to unpack and expose them.
Here's the first: "If the government taxes the rich to give money to the poor, that increases GDP."
A number of commentators took exception to my article about austerity earlier this week, in particular my claim that there is no robust economic theory according to which fiscal transfers increase GDP. Daniel Knowles of the Economist stated of me: "he alleges that transfers can't increase AD, which is totally wrong" and that "In theory, taxing people with a low marginal propensity to consume & giving more to those with a high MPC" increases GDP. (The "marginal propensity to consume" or MPC is the percentage of income that we consume as opposed to save.) This is a remarkably common thought, based on a confusion in the way that simple macroeconomics is often taught. It features in many political discussion, also, where it is suggested that cutting benefits "sucks demand out of the economy".
The naive idea behind this is the intuition that if people spend more of their money, rather than saving it, that will boost growth. Except that's just wrong. In orthodox macroeconomic models, money that isn't consumed is invested. And investment boosts GDP as well as consumption. So when the economy is operating properly, even if we could costlessly transfer money from high savers to high consumers and would not distort their incentives by doing so, there would be no effect upon GDP – it's just that a bit more of that GDP would be consumption and a bit less investment.
To get anywhere, therefore, this idea has to posit that the economy is subject to "over-saving", in the sense that some of what is saved is simply set aside for speculative or liquidity purposes and not actually invested into anything. We also have to posit that that extra ability to speculate and that extra liquidity isn't itself going to increase investment or consumption – the money is just (from the social point of view) pointlessly being not used (even if not using it might in some obscure way benefit the individual).
Well, spelling out how any of that could really work and why is pretty complicated and highly disputed, but there are indeed models (in particular, Keynesian models) in which it is assumed to occur. So let's park our objections and concede for now that there can be over-saving. Does that mean we can increase GDP by switching money from those that save a lot to those that consume a lot?
No! For at least three reasons. First, it isn't the amount of saving that counts here, but the amount of over-saving. The fact that one person saves a lot doesn't mean that she over-saves – it might be optimal for her to save a lot. And the fact that someone else hardly saves at all doesn't mean he doesn't over-save – ideally he might save even less! So transferring money from low MPC folk (say, "the rich") to high MPC folk (say, "the poor") might increase over-saving, rather than reducing it.
To make the concept of transfers from over-savers to non-over-savers work we'd need some systematic way to identify those that over-save and give money to those that do not over-save. Simply stating "I'll take it from an over-saver and give it to an optimal saver" isn't describing a policy – just an act. To make it a policy, we'd have to set out how we find these inefficient over-savers. And the government simply lacks any way to do that. It has no idea what the optimal level of saving even is, let alone who is saving at that optimal level and who is not.
The second problem is that even if it could find over-savers and efficient consumers, it cannot guarantee how they will react to funds being transferred between them. For example, an over-saver subject to a extra tax might react by saying "Golly! Things are even worse than I thought and the risk of unexpected demands on me even greater! I'd best save even more." Conversely, the person that saves very little might react to being given a transfer by saying "Finally! A bit of cash! But who knows when I might get more? I'll set this aside for a rainy day." In that case the transfer results in even more over-saving, because both the taxed and the beneficiary react to the transfer by increasing their over-saving.
Third, even if we could find over-savers and perfect savers, and even if the way they reacted to the transfer would not offset its effect, the GDP increase would still have to exceed the highly non-trivial costs of the process. The deadweight cost of taxation is estimated by the Treasury at around 30% of the value of the tax.
Transfers are done for what economists call "equity" reasons – they are there to help the poor or the sick and so on. That's good, but the process of doing it intrinsically distorts incentives, which if anything must reduce GDP. Whether you believe in a high benefits or low benefits society, it's just daft to believe that raising benefits will make the economy grow faster.
Here's the second fallacy: "When calculating the growth in GDP, we should net off the growth in debt."
That fallacy featured, for example, in the latest The Deep End's "Heresy of the Week" and is quite common in right-wing discussions about the state of the economy. The intuition behind the thought here goes rather like this. If you see me consuming more one month, does that mean I've got richer? If my consuming more is because my income has risen, then all well and good. But if I'm consuming more because I've taken out a large loan and am blowing it on short-term consumption, then I have really become richer. Again, suppose I buy a house – am I now richer? No, because I've had to part with money to buy the house. But suppose I didn't use my own money? Then we must net off any debts I took on to buy the house in assessing my ultimate net worth.
So if, say, GDP rises £1 but debts rise £5.40, the thought goes, we aren't really £1 better off at all, because we must later save up something extra – foregoing some consumption – to pay off the £5.40. Are we not, in fact, £4.40 worse off?
Well, no. That's just wrong. Let's see why. Suppose that in country ABC land there are three persons, A, B and C, who all work and produce £1000 of output. Then the GDP of ABC land is £3000. But suppose we discover that that A has borrowed £100 from B, B borrowed £100 from C, and C borrowed £100 from A. Then it turns out that as well as £3000 of output, ABC land has £300 of debt. Does that make us think the "true" GDP or wealth of ABC land is any less? No, because the debts of the ABC dwellers are all to each other. The collective debt of ABC to anyone else is zero.
Hence the relevant debt-related question about the wealth of a country is not how much debt its citizens owe to each other, but, rather, the net debts its citizens owe to or are owed by the rest of the world.
The fact that Britons increased their debts to each other during the 2000s may well be a reflection of their unrealistic expectations of the future, and might increase internal stresses in the financial system and damage the future growth in the economy. But it does not make the GDP accumulation at the time any less real. That's just wrong.