paradox of thrift

How much should we save?

pnrj critique of neoclassical economics, macroeconomics, public policy , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

JDN 2457215 EDT 15:43.

One of the most basic questions in macroeconomics has oddly enough received very little attention: How much should we save? What is the optimal level of saving?

At the microeconomic level, how much you should save basically depends on what you think your income will be in the future. If you have more income now than you think you’ll have later, you should save now to spend later. If you have less income now than you think you’ll have later, you should spend now and dissave—save negatively, otherwise known as borrowing—and pay it back later. The life-cycle hypothesis says that people save when they are young in order to retire when they are old—in its strongest form, it says that we keep our level of spending constant across our lifetime at a value equal to our average income. The strongest form is utterly ridiculous and disproven by even the most basic empirical evidence, so usually the hypothesis is studied in a weaker form that basically just says that people save when they are young and spend when they are old—and even that runs into some serious problems.

The biggest problem, I think, is that the interest rate you receive on savings is always vastly less than the interest rate you pay on borrowing, which in turn is related to the fact that people are credit-constrained—they generally would like to borrow more than they actually can. It also has a lot to do with the fact that our financial system is an oligopoly; banks make more profits if they can pay savers less and charge borrowers more, and by colluding with each other they can control enough of the market that no major competitors can seriously undercut them. (There is some competition, however, particularly from credit unions—and if you compare these two credit card offers from University of Michigan Credit Union at 8.99%/12.99% and Bank of America at 12.99%/22.99% respectively, you can see the oligopoly in action as the tiny competitor charges you a much fairer price than the oligopoly beast. 9% means doubling in just under eight years, 13% means doubling in a little over five years, and 23% means doubling in three years.) Another very big problem with the life-cycle theory is that human beings are astonishingly bad at predicting the future, and thus our expectations about our future income can vary wildly from the actual future income we end up receiving. People who are wise enough to know that they do not know generally save more than they think they’ll need, which is called precautionary saving. Combine that with our limited capacity for self-control, and I’m honestly not sure the life-cycle hypothesis is doing any work for us at all.

But okay, let’s suppose we had a theory of optimal individual saving. That would still leave open a much larger question, namely optimal aggregate saving. The amount of saving that is best for each individual may not be best for society as a whole, and it becomes a difficult policy challenge to provide incentives to make people save the amount that is best for society.

Or it would be, if we had the faintest idea what the optimal amount of saving for society is. There’s a very simple rule-of-thumb that a lot of economists use, often called the golden rule (not to be confused with the actual Golden Rule, though I guess the idea is that a social optimum is a moral optimum), which is that we should save exactly the same amount as the share of capital in income. If capital receives one third of income (This figure of one third has been called a “law”, but as with most “laws” in economics it’s really more like the Pirate Code; labor’s share of income varies across countries and years. I doubt you’ll be surprised to learn that it is falling around the world, meaning more income is going to capital owners and less is going to workers.), then one third of income should be saved to make more capital for next year.

When you hear that, you should be thinking: “Wait. Saved to make more capital? You mean invested to make more capital.” And this is the great sleight of hand in the neoclassical theory of economic growth: Saving and investment are made to be the same by definition. It’s called the savings-investment identity. As I talked about in an earlier post, the model seems to be that there is only one kind of good in the world, and you either use it up or save it to make more.

But of course that’s not actually how the world works; there are different kinds of goods, and if people stop buying tennis shoes that doesn’t automatically lead to more factories built to make tennis shoes—indeed, quite the opposite.If people reduce their spending, the products they no longer buy will now accumulate on shelves and the businesses that make those products will start downsizing their production. If people increase their spending, the products they now buy will fly off the shelves and the businesses that make them will expand their production to keep up.

In order to make the savings-investment identity true by definition, the definition of investment has to be changed. Inventory accumulation, products building up on shelves, is counted as “investment” when of course it is nothing of the sort. Inventory accumulation is a bad sign for an economy; indeed the time when we see the most inventory accumulation is right at the beginning of a recession.

As a result of this bizarre definition of “investment” and its equation with saving, we get the famous Paradox of Thrift, which does indeed sound paradoxical in its usual formulation: “A global increase in marginal propensity to save can result in a reduction in aggregate saving.” But if you strip out the jargon, it makes a lot more sense: “If people suddenly stop spending money, companies will stop investing, and the economy will grind to a halt.” There’s still a bit of feeling of paradox from the fact that we tried to save more money and ended up with less money, but that isn’t too hard to understand once you consider that if everyone else stops spending, where are you going to get your money from?

So what if something like this happens, we all try to save more and end up having no money? The government could print a bunch of money and give it to people to spend, and then we’d have money, right? Right. Exactly right, in fact. You now understand monetary policy better than most policymakers. Like a basic income, for many people it seems too simple to be true; but in a nutshell, that is Keynesian monetary policy. When spending falls and the economy slows down as a result, the government should respond by expanding the money supply so that people start spending again. In practice they usually expand the money supply by a really bizarre roundabout way, buying and selling bonds in open market operations in order to change the interest rate that banks charge each other for loans of reserves, the Fed funds rate, in the hopes that banks will change their actual lending interest rates and more people will be able to borrow, thus, ultimately, increasing the money supply (because, remember, banks don’t have the money they lend you—they create it).

We could actually just print some money and give it to people (or rather, change a bunch of numbers in an IRS database), but this is very unpopular, particularly among people like Ron Paul and other gold-bug Republicans who don’t understand how monetary policy works. So instead we try to obscure the printing of money behind a bizarre chain of activities, opening many more opportunities for failure: Chiefly, we can hit the zero lower bound where interest rates are zero and can’t go any lower (or can they?), or banks can be too stingy and decide not to lend, or people can be too risk-averse and decide not to borrow; and that’s not even to mention the redistribution of wealth that happens when all the money you print is given to banks. When that happens we turn to “unconventional monetary policy”, which basically just means that we get a little bit more honest about the fact that we’re printing money. (Even then you get articles like this one insisting that quantitative easing isn’t really printing money.)

I don’t know, maybe there’s actually some legitimate reason to do it this way—I do have to admit that when governments start openly printing money it often doesn’t end well. But really the question is why you’re printing money, whom you’re giving it to, and above all how much you are printing. Weimar Germany printed money to pay off odious war debts (because it totally makes sense to force a newly-established democratic government to pay the debts incurred by belligerent actions of the monarchy they replaced; surely one must repay one’s debts). Hungary printed money to pay for rebuilding after the devastation of World War 2. Zimbabwe printed money to pay for a war (I’m sensing a pattern here) and compensate for failed land reform policies. In all three cases the amount of money they printed was literally billions of times their original money supply. Yes, billions. They found their inflation cascading out of control and instead of stopping the printing, they printed even more. The United States has so far printed only about three times our original monetary base, still only about a third of our total money supply. (Monetary base is the part that the Federal reserve controls; the rest is created by banks. Typically 90% of our money is not monetary base.) Moreover, we did it for the right reasons—in response to deflation and depression. That is why, as Matthew O’Brien of The Atlantic put it so well, the US can never be Weimar.

I was supposed to be talking about saving and investment; why am I talking about money supply? Because investment is driven by the money supply. It’s not driven by saving, it’s driven by lending.

Now, part of the underlying theory was that lending and saving are supposed to be tied together, with money lent coming out of money saved; this is true if you assume that things are in a nice tidy equilibrium. But we never are, and frankly I’m not sure we’d want to be. In order to reach that equilibrium, we’d either need to have full-reserve banking, or banks would have to otherwise have their lending constrained by insufficient reserves; either way, we’d need to have a constant money supply. Any dollar that could be lent, would have to be lent, and the whole debt market would have to be entirely constrained by the availability of savings. You wouldn’t get denied for a loan because your credit rating is too low; you’d get denied for a loan because the bank would literally not have enough money available to lend you. Banking would have to be perfectly competitive, so if one bank can’t do it, no bank can. Interest rates would have to precisely match the supply and demand of money in the same way that prices are supposed to precisely match the supply and demand of products (and I think we all know how well that works out). This is why it’s such a big problem that most macroeconomic models literally do not include a financial sector. They simply assume that the financial sector is operating at such perfect efficiency that money in equals money out always and everywhere.

So, recognizing that saving and investment are in fact not equal, we now have two separate questions: What is the optimal rate of saving, and what is the optimal rate of investment? For saving, I think the question is almost meaningless; individuals should save according to their future income (since they’re so bad at predicting it, we might want to encourage people to save extra, as in programs like Save More Tomorrow), but the aggregate level of saving isn’t an important question. The important question is the aggregate level of investment, and for that, I think there are two ways of looking at it.

The first way is to go back to that original neoclassical growth model and realize it makes a lot more sense when the s term we called “saving” actually is a funny way of writing “investment”; in that case, perhaps we should indeed invest the same proportion of income as the income that goes to capital. An interesting, if draconian, way to do so would be to actually require this—all and only capital income may be used for business investment. Labor income must be used for other things, and capital income can’t be used for anything else. The days of yachts bought on stock options would be over forever—though so would the days of striking it rich by putting your paycheck into a tech stock. Due to the extreme restrictions on individual freedom, I don’t think we should actually do such a thing; but it’s an interesting thought that might lead to an actual policy worth considering.

But a second way that might actually be better—since even though the model makes more sense this way, it still has a number of serious flaws—is to think about what we might actually do in order to increase or decrease investment, and then consider the costs and benefits of each of those policies. The simplest case to analyze is if the government invests directly—and since the most important investments like infrastructure, education, and basic research are usually done this way, it’s definitely a useful example. How is the government going to fund this investment in, say, a nuclear fusion project? They have four basic ways: Cut spending somewhere else, raise taxes, print money, or issue debt. If you cut spending, the question is whether the spending you cut is more or less important than the investment you’re making. If you raise taxes, the question is whether the harm done by the tax (which is generally of two flavors; first there’s the direct effect of taking someone’s money so they can’t use it now, and second there’s the distortions created in the market that may make it less efficient) is outweighed by the new project. If you print money or issue debt, it’s a subtler question, since you are no longer pulling from any individual person or project but rather from the economy as a whole. Actually, if your economy has unused capacity as in a depression, you aren’t pulling from anywhere—you’re simply adding new value basically from thin air, which is why deficit spending in depressions is such a good idea. (More precisely, you’re putting resources to use that were otherwise going to lay fallow—to go back to my earlier example, the tennis shoes will no longer rest on the shelves.) But if you do not have sufficient unused capacity, you will get crowding-out; new debt will raise interest rates and make other investments more expensive, while printing money will cause inflation and make everything more expensive. So you need to weigh that cost against the benefit of your new investment and decide whether it’s worth it.

This second way is of course a lot more complicated, a lot messier, a lot more controversial. It would be a lot easier if we could just say: “The target investment rate should be 33% of GDP.” But even then the question would remain as to which investments to fund, and which consumption to pull from. The abstraction of simply dividing the economy into “consumption” versus “investment” leaves out matters of the utmost importance; Paul Allen’s 400-foot yacht and food stamps for children are both “consumption”, but taxing the former to pay for the latter seems not only justified but outright obligatory. The Bridge to Nowhere and the Humane Genome Project are both “investment”, but I think we all know which one had a higher return for human society. The neoclassical model basically assumes that the optimal choices for consumption and investment are decided automatically (automagically?) by the inscrutable churnings of the free market, but clearly that simply isn’t true.

In fact, it’s not always clear what exactly constitutes “consumption” versus “investment”, and the particulars of answering that question may distract us from answering the questions that actually matter. Is a refrigerator investment because it’s a machine you buy that sticks around and does useful things for you? Or is it consumption because consumers buy it and you use it for food? Is a car an investment because it’s vital to getting a job? Or is it consumption because you enjoy driving it? Someone could probably argue that the appreciation on Paul Allen’s yacht makes it an investment, for instance. Feeding children really is an investment, in their so-called “human capital” that will make them more productive for the rest of their lives. Part of the money that went to the Humane Genome Project surely paid some graduate student who then spent part of his paycheck on a keg of beer, which would make it consumption. And so on. The important question really isn’t “is this consumption or investment?” but “Is this worth doing?” And thus, the best answer to the question, “How much should we save?” may be: “Who cares?”

No, capital taxes should not be zero

pnrj critique of neoclassical economics, public policy , , , , , , , , , , , , , , , , , , ,

JDN 2456998 PST 11:38.

It’s an astonishingly common notion among neoclassical economists that we should never tax capital gains, and all taxes should fall upon labor income. Here Scott Sumner writing for The Economist has the audacity to declare this a ‘basic principle of economics’. Many of the arguments are based on rather esoteric theorems like the Atkinson-Stiglitz Theorem (I thought you were better than that, Stiglitz!) and the Chamley-Judd Theorem.

All of these theorems rest upon two very important assumptions, which many economists take for granted—yet which are utterly and totally untrue. For once it’s not assumed that we are infinite identical psychopaths; actually psychopaths might not give wealth to their children in inheritance, which would undermine the argument in a different way, by making each individual have a finite time horizon. No, the assumptions are that saving is the source of investment, and investment is the source of capital income.

Investment is the source of capital, that’s definitely true—the total amount of wealth in society is determined by investment. You do have to account for the fact that real investment isn’t just factories and machines, it’s also education, healthcare, infrastructure. With that in mind, yes, absolutely, the total amount of wealth is a function of the investment rate.

But that doesn’t mean that investment is the source of capital income—because in our present system the distribution of capital income is in no way determined by real investment or the actual production of goods. Virtually all capital income comes from financial markets, which are rife with corruption—they are indeed the main source of corruption that remains in First World nations—and driven primarily by arbitrage and speculation, not real investment. Contrary to popular belief and economic theory, the stock market does not fund corporations; corporations fund the stock market. It’s this bizarre game our society plays, in which a certain portion of the real output of our productive industries is siphoned off so that people who are already rich can gamble over it. Any theory of capital income which fails to take these facts into account is going to be fundamentally distorted.

The other assumption is that investment is savings, that the way capital increases is by labor income that isn’t spent on consumption. This isn’t even close to true, and I never understood why so many economists think it is. The notion seems to be that there is a certain amount of money in the world, and what you don’t spend on consumption goods you can instead spend on investment. But this is just flatly not true; the money supply is dynamically flexible, and the primary means by which money is created is through banks creating loans for the purpose of investment. It’s that I term I talked about in my post on the deficit; it seems to come out of nowhere, because that’s literally what happens.

On the reasoning that savings is just labor income that you don’t spend on consumption, then if you compute the figure W – C , wages and salaries minus consumption, that figure should be savings, and it should be equal to investment. Well, that figure is negative—for reasons I gave in that post. Total employee compensation in the US in 2014 is $9.2 trillion, while total personal consumption expenditure is $11.4 trillion. The reason we are able to save at all is because of government transfers, which account for $2.5 trillion. To fill up our GDP to its total of $16.8 trillion, you need to add capital income: proprietor income ($1.4 trillion) and receipts on assets ($2.1 trillion); then you need to add in the part of government spending that isn’t transfers ($1.4 trillion).

If you start with the fanciful assumption that the way capital increases is by people being “thrifty” and choosing to save a larger proportion of their income, then it makes some sense not to tax capital income. (Scott Sumner makes exactly that argument, having us compare two brothers with equal income, one of whom chooses to save more.) But this is so fundamentally removed from how capital—and for that matter capitalism—actually operates that I have difficulty understanding why anyone could think that it is true.

The best I can come up with is something like this: They model the world by imagining that there is only one good, peanuts, and everyone starts with the same number of peanuts, and everyone has a choice to either eat their peanuts or save and replant them. Then, the total production of peanuts in the future will be due to the proportion of peanuts that were replanted today, and the amount of peanuts each person has will be due to their past decisions to save rather than consume. Therefore savings will be equal to investment and investment will be the source of capital income.

I bet you can already see the problem even in this simple model, if we just relax the assumption of equal wealth endowments: Some people have a lot more peanuts than others. Why do some people eat all their peanuts? Well it probably has something to do with the fact they’d starve if they didn’t. Reducing your consumption below the level at which you can survive isn’t “thrifty”, it’s suicidal. (And if you think this is a strawman, the IMF has literally told Third World countries that their problem is they need to save more. Here they are arguing that in Ghana.) In fact, economic growth leads to saving, not the other way around. Most Americans aren’t starving, and could probably stand to save more than we do, but honestly it might not be good if we did—everyone trying to save more can lead to the Paradox of Thrift and cause a recession.

Even worse, in that model world, there is only capital income. There is no such thing as labor income, only the number of peanuts you grow from last year’s planting. If we now add in labor income, what happens? Well, peanuts don’t work anymore… let’s try robots. You have a certain number of robots, and you can either use the robots to do things you need (including somehow feeding you, I guess), or you can use them to build more robots to use later. You can also build more robots yourself. Then the “zero capital tax” argument amounts to saying that the government should take some of your robots for public use if you made them yourself, but not if they were made by other robots you already had.

In order for that argument to carry through, you need to say that there was no such thing as an initial capital endowment; all robots that exist were either made by their owners or saved from previous construction. If there is anyone who simply happened to be born with more robots, or has more because they stole them from someone else (or, more likely, both, they inherited from someone who stole), the argument falls apart.

And even then you need to think about the incentives: If capital income is really all from savings, then taxing capital income provides an incentive to spend. Is that a bad thing? I feel like it isn’t; the economy needs spending. In the robot toy model, we’re giving people a reason to use their robots to do actual stuff, instead of just leaving them to make more robots. That actually seems like it might be a good thing, doesn’t it? More stuff gets done that helps people, instead of just having vast warehouses full of robots building other robots in the hopes that someday we can finally use them for something. Whereas, taxing labor income may give people an incentive not to work, which is definitely going to reduce economic output. More precisely, higher taxes on labor would give low-wage workers an incentive to work less, and give high-wage workers an incentive to work more, which is a major part of the justification of progressive income taxes. A lot of the models intended to illustrate the Chamley-Judd Theorem assume that taxes have an effect on capital but no effect on labor, which is kind of begging the question.

Another thought that occurred to me is: What if the robots in the warehouse are all destroyed by a war or an earthquake? And indeed the possibility of sudden capital destruction would be a good reason not to put everything into investment. This is generally modeled as “uninsurable depreciation risk”, but come on; of course it’s uninsurable. All real risk is uninsurable in the aggregate. Insurance redistributes resources from those who have them but don’t need them to those who suddenly find they need them but don’t have them. This actually does reduce the real risk in utility, but it certainly doesn’t reduce the real risk in terms of goods. Stephen Colbert made this point very well: “Obamacare needs the premiums of healthier people to cover the costs of sicker people. It’s a devious con that can only be described as—insurance.” (This suggests that Stephen Colbert understands insurance better than many economists.) Someone has to make that new car that you bought using your insurance when you totaled the last one. Insurance companies cannot create cars or houses—or robots—out of thin air. And as Piketty and Saez point out, uninsurable risk undermines the Chamley-Judd Theorem. Unlike all these other economists, Piketty and Saez actually understand capital and inequality.
Sumner hand-waves that point away by saying we should just institute a one-time transfer of wealth to equalized the initial distribution, as though this were somehow a practically (not to mention politically) feasible alternative. Ultimately, yes, I’d like to see something like that happen; restore the balance and then begin anew with a just system. But that’s exceedingly difficult to do, while raising the tax rate on capital gains is very easy—and furthermore if we leave the current stock market and derivatives market in place, we will not have a just system by any stretch of the imagination. Perhaps if we can actually create a system where new wealth is really due to your own efforts, where there is no such thing as inheritance of riches (say a 100% estate tax above $1 million), no such thing as poverty (a basic income), no speculation or arbitrage, and financial markets that actually have a single real interest rate and offer all the credit that everyone needs, maybe then you can say that we should not tax capital income.

Until then, we should tax capital income, probably at least as much as we tax labor income.

Fear not the deficit

pnrj public policy , , , , , , , , , , , , , , , , , ,

JDN 2456984 PST 12:20.

The deficit! It’s big and scary! And our national debt is rising by the second, says a “debt clock” that is literally just linearly extrapolating the trend. You don’t actually think that there are economists marking down every single dollar the government spends and uploading it immediately, do you? We’ve got better things to do. Conservatives will froth at the mouth over how Obama is the “biggest government spender in world history“, which is true if you just look at the dollar amounts, but of course it is; Obama is the president of the richest country in world history. If the government continues to tax at the same rate and spend what it taxes, government spending will be a constant proportion of GDP (which isn’t quite true, but it’s pretty close; there are ups and downs but for the last 40 years or so federal spending is generally in the range 30% to 35% of GDP), and the GDP of the United States is huge, and far beyond that of any other nation not only today, but ever. This is particularly true if you use nominal dollars, but it’s even true if you use inflation-adjusted real GDP. No other nation even gets close to US GDP, which is about to reach $17 trillion a year (unless you count the whole European Union as a nation, in which case it’s a dead heat).

China recently passed us if you use purchasing-power-parity, but that really doesn’t mean much, because purchasing-power-parity, or PPP, is a measure of standard of living, not a measure of a nation’s total economic power. If you want to know how well people in a country live, you use GDP per capita (that is, per person) PPP. But if you want to know a country’s capacity to influence the world economy, what matters is so-called real GDP, which is adjusted for inflation and international exchange rates. The difference is that PPP will tell you how many apples a person can buy, but real GDP will tell you how many aircraft carriers a government can build. The US is still doing quite well in that department, thank you; we have 10 of the world’s 20 active aircraft carriers, which is to say as many as everyone else combined. The US has 4% of the world’s population and 24% of the world’s economic output.

In particular, GDP in the US has been growing rather steadily since the Great Recession, and we are now almost recovered from the Second Depression and back to our equilibrium level of unemployment and economic growth. As the economy grows, government spending grows alongside it. Obama has actually presided over a decrease in the proportion of government spending relative to GDP, largely because of all this political pressure to reduce the deficit and stop the growth of the national debt. Under Obama the deficit has dropped dramatically.

But what is the deficit, anyway? And how can the deficit be decreasing if the debt clock keeps ticking up?

The government deficit is simply the difference between total government spending and total government revenue. If the government spends $3.90 trillion and takes in $3.30 trillion, the deficit is going to be $0.60 trillion, or $600 billion. In the rare case that you take in more than you spend, the deficit would be negative; we call that a surplus instead. (This almost never happens.)

Because of the way the US government is financed, the deficit corresponds directly to the national debt, which is the sum of all outstanding loans to the government. Every time the government spends more than it takes in, it makes up the difference by taking out a loan, in the form of a Treasury bond. As long as the deficit is larger than zero, the debt will increase. Think of the debt as where you are, and the deficit as how fast you’re going; you can be slowing down, but you’ll continue to move forward as long as you have some forward momentum.

Who is giving us these loans? You can look at the distribution of bondholders here. About a third of the debt is owned by the federal government itself, which makes it a very bizarre notion of “debt” indeed. Of the rest, 21% is owned by states or the Federal Reserve, so that’s also a pretty weird kind of debt. Only 55% of the total debt is owned by the public, and of those 39% are people and corporations within the United States. That means that only 33% of the national debt is actually owned by foreign people, corporations, or governments. What we actually owe to China is about $1.4 trillion. That’s a lot of money (it’s literally enough to make an endowment that would end world hunger forever), but our total debt is almost $18 trillion, so that’s only 8%.

When most people see these huge figures they panic: “Oh my god, we owe $18 trillion! How will we ever repay that!” Well, first of all, our GDP is $17 trillion, so we only owe a little over one year of income. (I wish I only owed one year of income in student loans….)

But in fact we don’t really owe it at all, and we don’t need to ever repay it. Chop off everything that’s owned by US government institutions (including the Federal Reserve, which is only “quasi-governmental”), and the figure drops down to $9.9 trillion. If by we you mean American individuals and corporations, then obviously we don’t owe back the debt that’s owned by ourselves, so take that off; now you’re looking at $6 trillion. That’s only about 4 months of total US economic output, or less than two years of government revenue.

And it gets better! The government doesn’t need to somehow come up with that money; they don’t even have to raise it in taxes. They can print that money, because the US government has a sovereign currency and the authority to print as much as we want. Really, we have the sovereign currency, because the US dollar is the international reserve currency, the currency that other nations hold in order to make exchanges in foreign markets. Other countries buy our money because it’s a better store of value than their own. Much better, in fact; the US has the most stable inflation rate in the world, and has literally never undergone hyperinflation. Better yet, the last time we had prolonged deflation was the Great Depression. This system is self-perpetuating, because being the international reserve currency also stabilizes the value of your money.

This is why it’s so aggravating to me when people say things like “the government can’t afford that” or “the government is broke” or “that money needs to come from somewhere”. No, the government can’t be broke! No, the money doesn’t have to come from somewhere! The US government is the somewhere from which the world’s money comes. If there is one institution in the world that can never, ever be broke, it is the US government. This gives our government an incredible amount of power—combine that with our aforementioned enormous GDP and fleet of aircraft carriers, and you begin to see why the US is considered a global hegemon.

To be clear: I’m not suggesting we eliminate all taxes and just start printing money to pay for everything. Taxes are useful, and we should continue to have them—ideally we would make them more progressive than they presently are. But it’s important to understand why taxes are useful; it’s really not that they are “paying for” government services. It’s actually more that they are controlling the money supply. The government creates money by spending, then removes money by taxing; in this way we maintain a stable growth of the money supply that allows us to keep the economy running smoothly and maintain inflation at a comfortable level. Taxes also allow the government to redistribute income from those who have it and save it to those who need it and will spend it—which is all the more reason for them to be progressive. But in theory we could eliminate all taxes without eliminating government services; it’s just that this would cause a surge in inflation. It’s a bad idea, but by no means impossible.

When we have a deficit, the national debt increases. This is not a bad thing. This is a fundamental misconception that I hope to disabuse you of: Government debt is not like household debt or corporate debt. When people say things like “we need to stop spending outside our means” or “we shouldn’t put wars on the credit card”, they are displaying a fundamental misunderstanding of what government debt is. The government simply does not operate under the same kind of credit constraints as you and I.

First, the government controls its own interest rates, and they are always very low—typically the lowest in the entire economy. That already gives it a lot more power over its debt than you or I have over our own.

Second, the government has no reason to default, because they can always print more money. That’s probably why bondholders tolerate the fact that the government sets its own interest rates; sure, it only pays 0.5%, but it definitely pays that 0.5%.

Third, government debt plays a role in the functioning of global markets; one of the reasons why China is buying up so much of our debt is so that they can keep the dollar high in value and thus maintain their trade surplus. (This is why whenever someone says something like, “The government needs to stop going further into debt, just like how I tightened my belt and paid off my mortgage!” I generally reply, “So when was the last time someone bought your debt in order to prop up your currency?”) This is also why we can’t get rid of our trade deficit and maintain a “strong dollar” at the same time; anyone who wants to do that may feel patriotic, but they are literally talking nonsense. The stronger the dollar, the higher the trade deficit.

Fourth, as I already hinted at above, the government doesn’t actually need debt at all. Government debt, like taxation, is not actually a source of funding; it is a tool of monetary policy. (If you’re going to quote one sentence from this post, it should be the previous; that basically sums up what I’m saying.) Even without raising taxes or cutting spending, the government could choose not to issue bonds, and instead print cash. You could make a sophisticated economic argument for how this is really somehow “issuing debt with indefinite maturity at 0% interest”; okay, fine. But it’s not what most people think of when they think of debt. (In fact, sophisticated economic arguments can go quite the opposite way: there’s a professor at Harvard I may end up working with—if I get into Harvard for my PhD of course—who argues that the federal debt and deficit are literally meaningless because they can be set arbitrarily by policy. I think he goes too far, but I see his point.) This is why many economists were suggesting that in order to get around ridiculous debt-ceiling intransigence Obama could simply Mint the Coin.

Government bonds aren’t really for the benefit of the government, they’re for the benefit of society. They allow the government to ensure that there is always a perfectly safe investment that people can buy into which will anchor interest rates for the rest of the economy. If we ever did actually pay off all the Treasury bonds, the consequences could be disastrous.

Fifth, the government does not have a credit limit; they can always issue more debt (unless Congress is filled with idiots who won’t raise the debt ceiling!). The US government is the closest example in the world to what neoclassical economists call a perfect credit market. A perfect credit market is like an ideal rational agent; these sort of things only exist in an imaginary world of infinite identical psychopaths. A perfect credit market would have perfect information, zero transaction cost, zero default risk, and an unlimited quantity of debt; with no effort at all you could take out as much debt as you want and everyone would know that you are always guaranteed to pay it back. This is in most cases an utterly absurd notion—but in the case of the US government it’s actually pretty close.

Okay, now that I’ve deluged you with reasons why the national debt is fundamentally different from a household mortgage or corporate bond, let’s get back to talking about the deficit. As I mentioned earlier, the deficit is almost always positive; the government is almost always spending more money than it takes in. Most people think that is a bad thing; it is not.

It would be bad for a corporation to always run a deficit, because then it would never make a profit. But the government is not a for-profit corporation. It would be bad for an individual to always run a deficit, because eventually they would go bankrupt. But the government is not an individual.

In fact, the government running a deficit is necessary for both corporations to make profits and individuals to gain net wealth! The government is the reason why our monetary system is nonzero-sum.

This is actually so easy to see that most people who learn about it react with shock, assuming that it can’t be right. There can’t be some simple and uncontroversial equation that shows that government deficits are necessary for profits and savings. Actually, there is; and the fact that we don’t talk about this more should tell you something about the level of sophistication in our economic discourse.

Individuals do work, get paid wages W. (This also includes salaries and bonuses; it’s all forms of labor income.) They also get paid by government spending, G, and pay taxes, T. Let’s pretend that all taxing and spending goes to people and not corporations. This is pretty close to true, especially since corporations as big as Boeing frequently pay nothing in taxes. Corporate subsidies, while ridiculous, are also a small portion of spending—no credible estimate is above $300 billion a year, or less than 10% of the budget. (Without that assumption the equation has a couple more terms, but the basic argument doesn’t change.) People use their money to buy consumption goods, C. What they don’t spend they save, S.

S = (W + G – T) – C

I’m going to rearrange this for reasons that will soon become clear:

S = (W – C) + (G – T)

I’ll also subtract investment I from both sides, again for reasons that will become clear:

S – I = (W – C – I) + (G – T)

Corporations hire workers and pay them W. They make consumption goods which are then sold for C. They also sell to foreign companies and buy from foreign companies, exporting X and importing M. Since we have a trade deficit, this means that X < M. Finally, they receive investment I that comes in the form of banks creating money through loans (yes, banks can create money). Most of our monetary policy is in the form of trying to get banks to create more money by changing interest rates. Only when desperate do we actually create the money directly (I’m not sure why we do it this way). In any case, this yields a total net profit P.

P = C + I – W + (X – M)

Now, if the economy is functioning well, we want profits and savings to both be positive—both people and corporations will have more money on average next year then they had this year. This means that S > 0 and P > 0. We also don’t want the banks loaning out more money than people save—otherwise people go ever further into debt—so we actually want S > I, or S – I > 0. If S – I > 0, people are paying down their debts and gaining net wealth. If S – I < 0, people are going further into debt and losing net wealth. In a well-functioning economy we want people to be gaining net wealth.

In order to have P > 0, because X – M < 0 we need to have C + I > W. People have to spend more on consumption and investment than they are paid in wages—have to, absolutely have to, as a mathematical law—in order for corporations to make a profit.

But then if C + I > W, W – C – I < 0, which means that the first term of the savings equation is negative. In order for savings to be positive, it must be—again as a mathematical law—that G – T > 0, which means that government spending exceeds taxes. In order for both corporations to profit and individuals to save at the same time, the government must run a deficit.

There is one other way, actually, and that’s for X – M to be positive, meaning you run a trade surplus. But right now we don’t, and moreover, the world as a whole necessarily cannot. For the world as a whole, X = M. This will remain true at least until we colonize other planets. This means that in order for both corporate profits and individual savings to be positive worldwide, overall governments worldwide must spend more than they take in. It has to be that way, otherwise the equations simply don’t balance.

You can also look at it another way by adding the equations for S – I and P:

S – I + P = (G – T) + (X – M)

Finally, you can also derive this a third way. This is your total GDP which we usually call Y (“yield”, I think?); it’s equal to consumption plus investment plus government spending, plus net exports:

Y = C + I + G + (X – M)

It’s also equal to consumption plus profit plus saving plus taxes:

Y = C + P + S + T

So those two things must be the same:

C + S + T + P = C + I + G + (X – M)

Canceling and rearranging we get:

(S – I) + P = (G – T) + (X – M)

The sum of saving minus investment (which we can sort of think of as “net saving”) plus profit is equal to the sum of the government deficit and the trade surplus. (Usually you don’t see P in this sectoral balances equation because no distinction is made between consumers and corporations and P is absorbed into S.)

From the profit equation:

W = C + I + (X – M) – P

Put that back into our GDP equation:

Y = W + P + G

GDP is wages plus profits plus government spending.

That’s a lot of equations; simple equations, but yes, equations. Lots of people are scared by equations. So here, let me try to boil it down to a verbal argument. When people save and corporations make profits, money gets taken out of circulation. If no new money is added, the money supply will decrease as a result; this shrinks the economy (mathematically it must absolutely shrink it in nominal terms; theoretically it could cause deflation and not reduce real output, but in practice real output always goes down because deflation causes its own set of problems). New money can be created by banks, but the mechanism of creation requires that people go further into debt. This is unstable, and eventually people can’t borrow anymore and the whole financial system comes crashing down. The better way, then, is for the government to create new money. Yes, as we currently do things, this means the government will go further into debt; but that’s all right, because the government can continue to increase its debt indefinitely without having to worry about hitting a ceiling and making everything fall apart. We could also just print money instead, and in fact I think in many cases this is what we should do—but for whatever reason people always freak out when you suggest such a thing, invariably mentioning Zimbabwe. (And yes, Zimbabwe is in awful shape; but they didn’t just print money to cover a reasonable amount of deficit spending. They printed money to line their own pockets, and it was thousands of times more than what I’m suggesting. Also Zimbabwe has a much smaller economy; $1 trillion is 5% of US GDP, but it’s 8,000% of Zimbabwe’s. I’m suggesting we print maybe 4% of GDP; at the peak of the hyperinflation they printed something more like 100,000%.)

One last thing before I go. If investment suddenly drops, net saving will go up. If the government deficit and trade deficit remain constant, profits must go down. This drives firms into bankruptcy, driving wages down as well. This makes GDP fall—and you get a recession. A similar effect would occur if consumption suddenly drops. In both cases people will be trying to increase their net wealth, but in fact they won’t be able to—this is what’s called the paradox of thrift. You actually want to increase the government deficit under these circumstances, because then you will both add to GDP directly and allow profits and wages to go back up and raise GDP even further. Because GDP has gone down, tax income will go down, so if you insist on balancing the budget, you’ll cut spending and only make things worse.

Raising the government deficit generally increases economic growth. From these simple equations it looks like you could raise GDP indefinitely, but these are nominal figures—actual dollar amounts—so after a certain point all you’d be doing is creating inflation. Where exactly that point is depends on how your economy is performing relative to its potential capacity. In a recession you are far below capacity, so that’s just the time to spend. You’d only want a budget surplus if you actually thought you were above long-run capacity, because you’re depleting natural resources or causing too much inflation or something like that. And indeed, we hardly ever see budget surpluses.

So that, my dear reader, is why we don’t need to fear the deficit. Government debt is nothing like other forms of debt; profits and savings depend upon the government spending more than it takes in; deficits are highly beneficial during recessions; and the US government is actually in a unique position to never worry about defaulting on its debt.