(Note to readers: The first five paragraphs here are introductory, for readers new to economics. Analysis starts in paragraph 5, “So which is correct…”)
What should government do during a time of financial and economic crisis? Well, first, most people agree, the central bank should lower interest rates. The way this works in the U.S. is that the Federal Open Market Committee (FOMC), a committee of the Federal Reserve chaired by the Fed chairman, votes to set a target for new, lower interest-rates on U.S. Treasury securities. But the Fed can’t just issue declarations that ‘set’ those interest rates. Treasuries are bought and sold in competitive markets. So the Fed becomes a buyer — by purchasing lots of Treasuries, it increases the demand for Treasuries, which raises their price. A rising price of Treasuries (or any other bond) equates to a decreasing interest rate — i.e., if you could formerly buy a promise from the government to give you $100 in a year for $95, that means the interest rate was 5.27% ([100-95]/95); but if Federal Reserve bids up the price of that promise to 99 dollars, your interest rate is 1.01% ([100-99]/99). This is why the price of a bond and its interest rate always move in opposite directions — they’re just two ways of calculating the same phenomenon.
Why does lowering interest rates on Treasuries help? Since most financial assets in the U.S. are in some way linked to interest rates on Treasuries (i.e., the “risk-free rate”), a lower interest rate on Treasuries decreases interest rates throughout the economy — on your car loan, on your mortgage, on your business financing. This has to do with simple supply and demand — if there is no longer a supply of super-safe Treasuries that provide a yield of 5.27%, people with loanable funds will have to seek out other (riskier) options — like financing your house — to get that 5% return. So the Fed’s purchases of Treasuries, called “Open Market Operations,” lowers interest rates throughout the economy. (They also put more cash that can be lent to the private sector back into the hands of banks.) This makes it easier for businesses to get corporate financing to, say, expand their R&D facilities and hire more people to work in them. It makes it easier for you to afford a mortgage, which will make you more likely to buy a house, which, in turn, will be a boon to the construction industry, and its workers, who will consequently have more spending money, which they can spend on other industries that will also be boosted. So lower interest rates should bring more hiring and more spending, which should help the economy recover.
So why isn’t the Federal Reserve lowering interest rates right now? The trouble is, interest rates on U.S. Treasuries can’t really go any lower — they’re already near 0 in real terms (i.e, accounting for inflation). More, given that interest rates have been so long for so long, some fear that the good investment opportunities have already been undertaken — and sustained lower interest rates will just lead to easy money and cheap credit for unworthy enterprises, leading to mal-investment that could lead to another crash.
So what do we do now? Well, once the central bank is ‘out of ammo’ in terms of interest rates, economists disagree on what to do. Some — generally referred to as Keynesians — believe that the government itself should effect ‘fiscal stimulus.’ That means that the government suddenly takes on lots of new projects, like rebuilding national infrastructure, which would employ lots of workers, and put money in their pockets, which would help fight against the recession. Every dollar spent by the government, they argue, boosts the total economy by more than a dollar, since the workers the government employs will spend their money on haircuts and televisions that will also employ other people, who will then spend their money on…etc, etc. So government spending has a ‘multiplier effect.’ The Keynesians also highlight the fact that, because interest rates are so low right now, the government can issue debt to fund these projects especially cheaply.
But others — referred to as ‘austerians’ or advocates of ‘austerity’ — disagree. If you undertake large fiscal stimulus, the austerians argue, businesses will see the government’s growing debt and consequently anticipate future tax increases that, in the future, will inevitably cut into the profits of projects they were considering starting. So they’ll invest and expand less. Plus, they argue that the ‘multiplier effect’ is overstated. In the chaotic political process, special interests will ‘capture’ the benefits of any stimulus program. Instead, governments should cut taxes and concordantly cut spending, to restore business’s confidence that the country will be economically stable and hospitable to business into the future, and interest rates will be kept low. With that confidence, banks will lend more, and businesses will invest and expand and spend more. The debate between these two camps is, to put it mildly, voluminous and heated. Since the 2008 financial crisis, the Keynesian view has gained popularity among the economic commentariat, but countries have varied in their actions — China and the U.S. undertook large stimulus funding, while the U.K. has generally leaned more toward austerity.
So which is correct, stimulus or austerity? Well, as above, each position seems to have a logic to it; each seems reasonable in theory, in writing. So let’s ask another question — how would we go about testing each thesis? Well, we could find countries that have experienced crises, we could find out whether they implemented stimulus or austerity, and consequently see how they did. From there, we could figure out which fare better — countries under austerity or countries under stimulus. And this is the sort of debate we see on editorial pages every week — there’s a new economic report from one country or another, and the economic commentariat pounces upon it as proof of what they knew all along. But this isn’t exactly scientific. Republicans in the United States have argued that the disappointing economic recovery is proof that the Democrats’ 2009 stimulus did not work. But Obama’s defenders point out that we need to consider the counterfactual: They argue that economy would be even worse, much worse, without the stimulus, and so the stimulus ‘worked’ in bringing us from ‘unthinkably horrible’ to ‘disappointing’ economic conditions. Some, such as Paul Krugman, make the argument that the rigorous economic theory centering on the ‘multiplier effect’ demanded much more stimulus than Congress passed in 2009, and that only a full stimulus would really work. So the U.S.’s difficulties are no proof that stimulus doesn’t work. What about austerity? Many countries that have implemented austerity have not consequently seen robust growth, and some have even experienced new crises (whether Greece actually effected ‘austerity’ is debatable; in 2001, Argentina’s attempts at austerity prompted a public backlash so extreme the instability and loss of confidence caused a financial crisis). Can we conclude that austerity is harmful? Not quite. We need, again, to consider counterfactuals. Coming to either conclusion from these ‘data’ would be like concluding that, since death rates are higher among people who have gone to an emergency room lately than among the general population, emergency rooms must kill people.
So, again, how do we test the competing theses? Let’s make it even easier: How do we go about testing just the stimulus thesis alone? There are two difficulties: (1) Any time that a government enacted stimulus in response to a crisis, we’ll always have the counterfactual question, “But couldn’t things have been even worse without it?” (2) Modern states and hence modern fiscal policy are relatively novel — so our ‘sample’ of fiscal stimuli isn’t very deep. In economics we can’t rerun experiments multiple times, adjusting key variables in isolation and observing the effects, as we can in, say, chemistry. So how can we attribute causation?
The best we can do is what economists call a natural experiment — some historical event that ‘isolated’ the variable ‘stimulus’ from the context of ‘recession,’ so we can observe the former variable without contamination. That could be some set of circumstances in economically normal times that made policy-makers enact policy changes that were equivalent to a stimulus, but that weren’t a response to a recession. For advocates of the stimulus thesis, such as Paul Krugman, that natural experiment is WWII. From the stock-market crash of 1929, through the Great Depression, up until the U.S. began gearing up for WWII, economic growth in the U.S. was anemic or negative. In 1940, the U.S. started building new bases and munitions and planes, etc., in response to the possibility of war, not due to an economic downturn — and from there onward, for a long stretch, the economy took off. So when the variable ‘stimulus’ was isolated from the context of an already-advancing recession, it worked, and, according to the economists’ calculations, had a very large multiplier effect indeed.
Let me be clear that I take this argument extremely seriously. The case is compelling. In response to it, I want to call into question not this particular economic argument, but economic methodology — and hence, economics — as a whole.
Allow me me to draw a comparison to one of my other pet preoccupations: exercise physiology. I love to run; this morning, I ran a 5-k race. I ran my hardest and did decently well (17:10), but I could have potentially gone slower or faster. Why do I think that? Well, my pace varied throughout the race in ways that don’t seem to be explained purely by ability — at mile 2 I was running faster with less effort than I had been at mile 1.5, even though I wasn’t yet inspired by the site of the finish line, and even thought my muscles were more worn and soaked in lactic acid by that point. What can explain the difference? Not exercise physiology. Exercise physiology can give me a lot of very good basic guidelines to becoming a better runner — it can tell me what kinds of nutrients I need in my body, it can explain why I should mix fast track workouts on some days with long, ‘aerobic’ runs on others. But it can explain very little about why I felt I could run faster with less effort at mile 2. What ‘s the actual explanation? As far as I can tell, I sometimes have surges of adrenaline, where I feel motivated to catch one person ahead of me out of pride; while other times my adrenaline falls, without my control, as I get distracted by thoughts of other obligations and begin to think that the race doesn’t really matter. I feel the way I did at mile 2 when some inspiring thought wiggles its way into my head, through no action of my own. To stretch the metaphor, we could say that exercise physiology can describe my ‘potential output,’ and that potential output constrains how fast I can run, but my actual pace at any moment has to do with my motivation, my identification with my task, with choruses replaying in my head that bring goosebumps and adrenaline, what you might call as a catch-all, ‘spirit’ — a term that evokes ‘animal spirits’ but goes beyond just financial swings, and also connotes human choice, agency, and motivation, more generally.
My point here is that WWII probably had a lot of effects on ‘spirit.’ In fact, only an economist would really think of WWII as a clinically isolated boost in government expenditures, whose effects would therefore be replicable in any other context. American pride was pricked as never before by the attack on Pearl Harbor; millions of men were sent overseas, so those who stayed home no doubt felt they needed a really good reason why; families at home read harrowing letters from their soldiers; America as whole thought, correctly, that the fate of the free world depended upon it; Americans saw, everywhere, government propaganda that reminded them, correctly, how much their work and productivity mattered; and toward the end of the war, America felt itself emerging as the new global superpower.
My thesis is that all of this could amount to, in effect, a big national adrenaline surge — a shift in spirits that made people suddenly identify with their work in billions of small ways that, in aggregate, had enormous effects on GDP. Workers became more productive, business owners suddenly became more ambitious, because, well, suddenly, it really mattered. I don’t have any data on national adrenaline levels; I can’t run any regressions on any of these variables or their effects; but I can’t believe they don’t matter, because, well, I’m human.
And what this suggests is that it’s really implausible to imagine that we can isolate the effects of ‘government purchases’ from the context of “The Greatest, Goodest War Ever Fought, From Which our Heroic Boys Emerged Victorious.” Have I demolished the case against stimulus? Hardly. I think the logic of the case for stimulus is compelling; but I think the most widely-cited natural experiment is neither. And this problem would seem to contaminate empirical economics research more generally: Just as my running speed depends on my mysterious state of feeling ‘psyched up,’ my productivity at work varies with mysterious states of alertness, interest, and enthusiasm about the idea of being a hard worker. Since macro-economies are built on micro-economic behaviors, it really could be the case that the reason one nation is wealthier than another is actually explained by how psyched up and enthused about hard work the people in each country are. But since we can’t measure these, the regressions economists run will always attribute the difference to some other, measurable variable.
This sort of ties in to an interesting, telling, surprise in development economics recently, that came through a paper by a few Dutch economists. For the past many years, development economists have focused on “Randomized Control Trials” to test their theories of development. They find instances in which a form of aid intervention is distributed across a population completely randomly — that way, they can isolate the effect of the intervention itself. Consider the alternative: If we observed a bunch of localities, some of whose leaders required all children to go to school, and others’ who didn’t, and then saw that the the former localities were wealthier by X amount, we couldn’t conclude that all of X was explained by the education policy difference. Rather, it seems likely that the local leaders who required education would be more enlightened in other ways, too, that could explain most of the difference. Likewise, we can’t just observe the impact of a World Bank program that distributes aid non-randomly — because, then, it’s likely that the people who received the aid might’ve been selected for their exceptional dire straits, or because they were more politically well-connected, both of which would bias the sample. So development economists more and more require “Randomized Control Trials” (RCTs), which are frequently compared to ‘double-blind’ experiments in medicine, to provide proofs of the efficacy of any aid intervention.
But the problem, as the paper argues, is that RCTs are not actually double-blind. The objects know they are receiving aid, and may change their behavior accordingly. There could be an ‘aid placebo effect.’ And, indeed, the economists find statistical evidence of just that: When groups were randomly selected, and some were provided with ‘modern cowpea seeds’ while others were provided with traditional cowpea seeds, the former groups with the modern seeds had yields about 20% greater than the others. Amazing! But, more amazingly, this effect completely disappeared when the group with the modern cowpea seeds was not told they were getting modern seeds. In other words, it appears that the whole difference could be accounted for by behavioral changes — by farmers who worked harder and more optimistically at the thought that they had special, advanced seeds that would bring them enormous advantages.
In short, I’m joining those who critique the modern economics profession for becoming overly quantitative, while insufficiently nuanced and open in its approach to human behavior. It seems strange that economic theorists spend so much time making extraordinarily mathematically sophisticated extrapolations of the assumptions of microeconomic theory, when ordinary people would report that those assumptions are clearly untrue. It seems strange, too, that the Dutch study above caught so many (me included) by surprise — have we seriously never noticed, in our own work lives, how we all work harder and better and more productively (our yields increase, as it were), when we feel optimistic about the end result? How could we not assume the same of poor people who are told “these are modern seeds.”
(For those who are interested, my actual (currently completely unqualified) views on fiscal stimulus are: What fiscal stimulus? On the theory side of things, I lean more toward the stimulus than the austerian view, because I think the austerian assumptions about human behavior are unrealistic — i.e., I really doubt, and the evidence doubts, people fully take account of potential future tax hikes and decrease their expenditures to match the government’s increases. So an ideal stimulus could rescue the economy from a downward spiral of declining spending, declining profits, declining confidence, declining investment, job losses, and more declining spending ad infinitmum. So I believe the theory that supports an ideal stimulus. But in practice, stimulus is rarely implemented ideally or effectively: Government can’t just execute an order that boosts demand for the economy as a whole. It has to procure funding for various agencies that have their own, independent motives. Indeed, a very large portion of the 2009 ‘stimulus’ money was actually not spent in an even moderately timely manner, because the people who had final control over its expenditure had their own goals and interests aside from simply an aggregate demand boost. And then, stimulus has a lot of bad second-order effects: Every ‘temporary’ government program creates constituents who will demand that its expenditures be continued even long after the ‘stimulus’ justification has disappeared — planting the seeds of a new crisis, years on, driven by debt and loss of competitiveness. At the same time, I think it’s likely a bad idea to take austerity to the point of laying off government workers during an already-declining economy, which could add to a downward spiral. The correct position, then, seems to me to involve a mix of boosting demand through more aggressive deficit-financed tax cuts, having counter-cyclical social insurance policies in place, and using anomalously-low interest rates to undertake public-infrastructure projects that actually need to be done anyways and actually will be temporary. More importantly, politicians should take advantage of crises to undermine incumbent interests groups and cartels that drag on innovation and growth: A time when people are intensely worried about their pocketbooks is the right time to tell them that their local friendly doctor, yes him, is driving up medical costs by blocking reforms that would let nurses take on simpler tasks and reforms that would make some really simple and harmless medications available over the counter; or that their haircuts are so expensive partly because their state government got the idiotic idea that barbers need to be licensed. These kinds of growth-spurring reforms tend to get left ouf of the macroeconomic debate.)