Unfortunately, in between my fun (but demanding) day job and studying for the CFA, I am finding that I just don’t have the time and mental energy to sustain the level of blogging seen here recently. From now on, barring exigent circumstances I will be publishing two items weekly: one proper post, and another with links and some brief thoughts on them. It’s something I have been considering for a while, but the change was prompted by my hasty adventures this week over reserve requirements, which are somewhat symptomatic of my mental overstretch (I will return to the question in future, but I need to put some more serious thought in first). So I’m going to pull back a bit and focus more intently on what I see as my comparative advantage – conceptual/philisophical issues in economics. I hope this will give me time to interact more fully in the comments section and to phrase what I write more carefully. It has been amazing to pull in some fantastic commenters in the last few weeks; you guys are awesome and I really would love it if you stick around (the ‘follow by email’ and RSS feed buttons are on the top right!). I hope this change means I will be at my best with everything I write in future.
To whet your appetite, some things I have bouncing around my head at the moment include:
- Alternatives to the fiduciary theory of the firm (and towards a more general theory, including non-profit orgs)
- Some stuff on the economic concept of ‘capital’ (of which I am not a fan)
- The case for using market forecasts in public policy (esp. monetary)
- The role GDP accounting may be playing in the debate over the existence of ‘liquidity traps’
I may also re-publish (with light edits) some of the best posts I wrote in the period before most of you started tuning in.
Many thanks for reading,
Yesterday, some guy obviously *pretending* to be me wrote some seriously crazy stuff on my blog. He thought that reserve ratios made no difference to the banking system, and basically bought the MMT line that the only thing that matters is bank capital. But here’s a pretty simple reductio: if reserve requirements are 100%, then deposits finance zero lending. Since banks make money by lending, no bank will (voluntarily) pay interest on deposits. 100% reserve requirements completely divorce lending and deposits. Duh.
How could that guy be so wrong? Well, maybe he thought about it this way. Suppose reserve requirements were increased in such a way that there were no longer enough reserves to sustain the current level of deposits in the economy. In order to prevent a likely deflationary contraction (decrease in V), the central bank creates more reserves (increase in M). As long as no one changes their intertemporal consumption preferences or the balance of their portfolios (a whopping great big howler of an assumption), this *appears* to have no effect at all on the economy or the banking system. The Fed can just create more reserves, whilst managing the marginal cost of loanable funds, and everything seems to go along as normal. But this cannot be right, because if reserve ratios increase then deposits ‘fund’ less lending and banks won’t pay as much to compete for them. The central bank makes up the difference. Interest on deposits fall, and ends up being paid to the central bank instead (as I’m assuming for the moment the Fed funds rate stays the same, and this is supported by repo transactions which earn the Fed interest). Of course, this isn’t even remotely close to an equilibrium, as a reduction in the interest rate on deposits will cause depositors to re-balance their portfolios, affecting both AD and the natural rate of interest (and hence the optimal Fed funds rate under the policy regime).
Basically, an increase in the reserve ratio (under a monetary policy regime like in the US) redistributes seigniorage away from depositors and towards the central bank*. Depositors then take some action to mitigate this by re-balancing their portfolios, with the central bank then counteracting any effects of said re-balancing on the policy target. If any of this re-balancing is towards higher consumption, the Fed funds rate will have to increase (reflecting the change in intertemporal consumption preferences, and hence the natural rate of interest). That being said, as the Fed remits any net income to the Treasury, it’s a little more complicated as less taxation/bonds would be required to finance any given level of government spending.
In conclusion, that guy who took over my blog really should have just spent Sunday morning watching the tennis instead. What an idiot**.
*This, I think, is just another way of making Nick Rowe’s point in the comments on the previous post. I’m seriously considering just outsourcing all my beliefs to him, and being done with this whole ‘having opinions of my own’ thing
**In that guy’s defense, this stuff is (in my view) pretty damn hard. It’s unbelievably easy to make silly mistakes – even if you have a Nobel prize. So be nice… if you happen to see him around
Why do we have fractional reserve banking in a fiat currency world? (speculative, blue-sky thinking, tell me why I’m crazy)
UPDATE: I have a new post here, arguing that this post is wrong.
If the coin be locked up in chests, it is the same thing with regard to prices, as if it were annihilated – David Hume
OK guys, you’re going to have to bear with me on this one because I’m in blue-sky thinking mode – so usual caveats on speculative thinking apply double for this one. What would happen if the central bank just said “We’re implementing a 100% reserve ratio, but don’t worry we’re just going to create all those reserves in your account with us at a click of a button”? I can’t see how anything would happen at all. It wouldn’t be inflationary, because it would make no difference to people’s decisions or ability to use their money to chase goods and services – ‘the coin be locked up in chests’. But it would mean there are no bank runs, because ‘the money is always there’.
So, why can’t we just do this? The central bank can control the money supply as before to maintain a certain policy target, the only difference is there are now tons of new reserves sitting in accounts with the central bank that aren’t allowed to ‘go anywhere’ due to the 100% reserve requirement. The new reserves are created by the central bank as needed, and everything just goes on as before. Except that old-fashioned bank runs would not happen.
Am I crazy? What else would happen to the banking sector if we did this? How would it affect the natural rate of interest? This smacks of WAY-too-good-to-be-true, and I thought the most efficient way for me to find out where I’ve gone wrong here is to open myself up to public humiliation for having missed something totally obvious.
FWIW, the way I ended up at this conclusion was wondering as to what would happen if we increased or decreased reserve requirements, which seemed to me to be precisely nothing – as the central bank (if it was doing it’s job) would just offset the deflationary/inflationary effects caused by a massive reduction/increase of supply in the market for excess reserves. So long as the central bank controls the price in that market, then I can’t see why reserve requirements matter at all. Except that less reserves means more bank runs.
This is by way of saying that I can’t think of any objections to the full-reserve position that are not general objections to the marginal cost of funds being the key instrument of monetary policy, which would apply no matter the regulatory reserve requirement ratio.
- Nassim Taleb talking with Russ Roberts about “Antifragility”, his book-in-progress. I find Taleb’s style really frustrating, and I think he too often relies in informal argument on the kind of ‘just-so stories’ he has so rightly criticised in both Fooled by Randomness and Black Swan. But his intellect is incredibly fertile (Fooled by Randomness had a major influence on my intellectual development as a teenager), and Russ’ clear questions and follow-ups are the perfect foil to Taleb’s sometimes opaque way of speaking (although his writing, in my view, is very clear). And I can’t wait for the book.
- Waleed Hussain (of Wharton) has a working paper, entitled ‘An Alternative to the Fiduciary Theory of the Corporation‘ (H/T Matt Yglesias) which is so brilliant and raises so many complex and important issues that it made want to just drop everything I’m doing so I can think about it, work through it and tease out the implications. Unfortunately that isn’t an option, but expect posts on this topic in the not too distant future.
This Timothy Taylor post from last week on Gary Gorton’s very important and fascinating work on the financial crisis reminded me to get back to pursuing a line of thought I started a while ago, about a potentially fruitful analogy between what structured finance products accomplish on the micro level and what the financial sector does at the macro level*. I haven’t come close to thinking this all the way through yet, but I find thinking out loud very helpful sometimes. So please take this in the open spirit of inquiry in which it is offered.
First, Gorton. He has a theory of financial crises that relies on the distinction between information-sensitive and information-insensitive debt. His thought (in its barest form) is that people often have a need for financial assets that they don’t need to worry about how new information affects the possibility of a loss. Then something happens to make those assets information-sensitive, and people sell those assets in droves to reacquire information-insensitive assets. This then wreaks havoc in the repo market, with haircuts on short term secured lending increasing dramatically due to the uncertainty over the pricing of those assets – effectively leading to a net withdrawal from the banking system:
But how do we create safe assets in the first place? The economy is mostly made up of people doing risky things: starting businesses, investing in new factories, opening new divisions or product lines etc. There’s two basic ways you can turn risky things into safe things: diversify, and structure the cashflow pooled through said diversification. In my mind, I have the whole economy as looking like a giant securitisation vehicle, spitting out bank deposits / money market funds at the top and good ol’ equity at the bottom.
Now, how much information-insensitive debt can be produced from a given set of underlying assets is a basically a function of a) the individual riskiness of the assets and b) the correlation of the risks between those assets. (I intend at some point to build a toy model to explain and show this better, but just accept this on epistemological credit for now). If the underlying assets become more risky, or if the correlation between them increases, then that means a reduction in the amount of super-safe, information-insensitive assets that can be spit out the other end.
Now, I wonder what could possibly cause both a) an introduction of risk across the entire economy and b) correlation in those risks. Well, how about a fall in aggregate demand caused by an increase in the demand for money not offset by an increase in the supply of money by the central bank? That (if the Market Monetarists are right) introduces new risks to businesses all over the place, and the risk is highly correlated across the economy. The existing structure of the economy’s cash flows can no longer sustain the same level of super-safe assets, so therefore some previously information-insensitive assets have to become sensitive and you get your crisis a la Gorton.
So when people are talking about a ‘shortage of safe assets’, you either need to restructure the cashflows of the economy to have more loss-absorbency at the bottom to support more quality at the top, or solve the problem of the underlying increase in correlated risk. If this is right, then this seems to me a pretty powerful (but embryotic) theory of how an AD shortfall could cause (or seriously exacerbate) a financial crisis like the one we experienced.
Other assorted thoughts:
- I think this highlights a key problem with Gorton’s solution to the problem (massive extensions of deposit insurance), which is that it doesn’t in any way do anything to change the amount of super-safe assets the underlying economy can *actually* sustain
- All the above is not to invalidate all the other potential factors that led to the crisis, (and the chart above seems to show that the run on the repo market was already beginning by early 2008) but to identify the mechanism whereby it suddenly got a whole lot worse when NGDP (and expectations thereof) fell in mid-summer 2008, and why people are currently saying ‘we still have a shortage of safe assets’
- Felix Salmon’s question ‘Is information-insensitive debt a good thing‘, is one worth considering very seriously
* I’m very indebted to Arnold Kling, whose mantra “The nonfinancial sector wants to issue risky long-term liabilities and to hold safe short-term assets. The financial sector accommodates this by doing the opposite” first got me thinking down this line
[UPDATE: it does appear from reading the comments sections on a number of posts on this debate that pretty much everyone has accepted Nick Rowe's counter-example to the identity claim propounded by Dean Baker and Paul Krugman. However, it seems to me that most people, even if they have accepted the counter-example, have not yet seen the implications of it. That is what this post is (mostly) about.]
Ok, let’s try a slightly different way of thinking about this. Suppose the government wants to spend more money on x, where x could be anything. Let’s think about what happens if the government finances x through taxation, or through bonds. Assume for the sake of argument that the spending on and financing costs of x are evenly distributed through the current population, and that there are no incentive effects changing total output. Whether x is paid for through debt or taxation, money is initially just moved around the current population. The difference in the case of bonds is that everyone is left with an asset (bonds) and a corresponding liability (being part of the tax base), which initially cancel each other out.
Then, the next cohort is born and become part of the tax base. The older cohort still hold the same amount of bonds, but the liability is now spread out between both cohorts. So the older cohort now have a net asset, even though they haven’t had to forego any consumption to obtain it, and the younger cohort a net liability. If the younger cohort purchase the bonds from the older cohort, all does not end up equal – because the older cohort did not have to forego consumption to end up with bonds, whereas the younger cohort did.
So an increase in public debt does impose a burden on future cohorts (relative to tax financing), unless either a) there is a corresponding increase in bequests (i.e. assets the younger cohorts did not have to forego consumption in order to obtain) or b) the nominal interest rate on the debt is less than the rate of NGDP growth, in which case the debt can be rolled over to future cohorts forever without a tax ever being instituted (Samuelson 1958, H/T Nick Rowe). If it was the case that b) was true and a) was false, the first cohort have still increased their consumption from what it otherwise would have been, just not at anyone’s expense (although future cohort’s consumption would shift forwards in time, because they purchase bonds from the older cohort allowing that cohort to consume, but then do the same trick to the next cohort etc.).
But if you think a) is true, then whether or not the debt is held domestically or internationally is irrelevant, because it’s the increase in bequests to offset the future cohort’s liability that makes the difference no matter who holds the asset. I see no reason to think that marginal changes in bequests, insofar as they happen, would depend on whether the marginal debt was domestic or international (it would involve equal foregoing of consumption on the part of the older cohort either way). Welcome to Planet Landsburg*.
The whole debate boils down to this: unless you believe that corresponding marginal changes to voluntary bequests occur in response to marginal changes in public debt, deciding what mix of debt and taxation to plump for will have intergenerational distribution effects. Of course, it could be entirely appropriate that future people should foot some of the bill for the schools we build for them. But that is what would happen if we were to finance government spending with debt rather than taxes (unless you believe in something approximating Ricardian equivalence, or unless the debt is ponzi-sustainable a la Samuelson 1958).
[Note: This bit is more speculative, just throwing it out there] Furthermore, if we relax the assumption around the distribution of the tax burden to something more approximating reality, I think the intergenerational distribution effect actually becomes even more pronounced. This is because people living off their assets have lower taxes, and therefore form a relatively smaller part of the tax base. So when older cohorts are winding down their assets (i.e. selling bonds) in retirement, they are actually in a substantially better position (and the younger cohort in a worse position) than if you assume even distribution of taxation.
*For the record, I don’t (as an empirical matter) believe we live on Planet Landsburg, but would be delighted to be proven wrong
‘Take the biscuit’ – British idiom meaning “to be particularly bad, objectionable, or egregious” (Wiktionary)
Like Bob Murphy, I’m almost rendered speechless by Paul Krugman’s new post on debt*. The whole economics blogosphere went through this great big massive argument about Krugman and Dean Baker‘s claim that debt cannot by definition be a burden to future generations. Almost anyone who is an anybody wrote about it, and a whole lot of nobodies as well (including myself). A lot happened in that time. Nick Rowe showed very clearly that the apparent identity claim they were making was wrong, and Bob wrote a whole semi-Socratic dialogue on the subject, with like 12 different parts (and enough wit to fill a full sitcom run on NBC**). For those of you who missed out, here’s a quick recap of the arguments. Krugman and Baker make the observation that debt is money someone owes to someone else. It is an asset as well as a liability. So, as long as debt is held domestically, it doesn’t make any sense to say that debt can be a burden to future generations, because it’s just some of them paying money to some of the others. Within a closed system, it’s a tautology to say that the debt cannot be a burden to the people within the system, because the assets equal the liabilities.
But when you introduce overlapping generations, this identity breaks down when we make the distinction between cohorts and time periods. Here is my version of Nick’s argument, which I posted last week
Assume any person alive produces 100 apples a year. In year 1, Annie is the only person alive. The government says ’100 free apples for Annie’! This is financed by the government borrowing Annie’s apples, promising to give her 110 next year (10% interest). Annie produces and consumes 100 apples.
At the beginning of year 2, Annie gives berth to Bessie, and there’s been a 10% increase in apple productivity to 110 apples. Annie and Bessie each produce 110 apples. Annie’s bond matures and needs to be paid back. The government finances the 110 apples it owes to Annie by issuing a bond bought by Bessie. In year 2, Annie eats 220 apples, Bessie none. At the end of year 2, Annie dies of old age.
In year 3, Colin is born to Bessie, and apple productivity increase by 10% to 121 apples each. Bessie’s bond matures, but since Colin is a man and can’t give birth to anyone, and Bessie will die at the end of the year, there’s no market for a bond as there’s no prospect for it being paid back. Therefore, the government taxes Colin and takes his 121 apples and gives them to Bessie. Colin gets no apples, and Bessie gets 242. Bessie dies at the end of the year.
In year 4, apple productivity increases by 10%. Colin produces and consumes 133.1 apples and dies at the end of the year.
Had the government not issued any bonds, Annie would have eaten 210 apples rather than 320. Bessie would have eaten 231 instead of 242, and Colin would have gotten 254.1 apples rather than 133.1. The debt accomplishes a transfer of consumption from Colin to Annie even though they weren’t alive at the same time, and even though (in any given year) the output was exactly the same and was consumed only by people alive at that time.
When you factor in multiple overlapping generations, it can too be the case that no output is transferred between time periods, but future cohorts lose out. In my example, Annie really does eat apples at Colin’s expense. The only way that this wouldn’t be the case is if the bonds were voluntarily bequeathed from Annie to Bessie and Bessie to Colin, rather than them purchasing the bonds. Which is basically assuming Ricardian equivalence, an assumption which Krugman never made and believes is dubious anyway. Game, set and match.
Krugman didn’t respond at all to this line of argument. And that’s OK, if a little annoying as he was responsible for the whole debate in the first place (not that I’m complaining, I initially would have been in full agreement with him until Nick Rowe showed me the error of my ways). But then yesterday he criticized an op-ed which plainly misunderstood his original argument. Now I would be the first to say that Krugman was obviously well within his rights to point out that the piece completely misunderstood his argument, but he then merely repeated the original assertion, as if nothing else had happened in the interim!
What I was actually saying, of course, is that debt is a liability that we pass to the next generation — but it’s also an asset that we pass to the next generation.
No. This only works if you don’t have overlapping generations, as Nick’s/Bob’s/[my] examples clearly demonstrate. In order for the debt to be harmlessly ‘passed’ to the next generation, it requires an actual voluntary bequest, with the older generation deliberately foregoing consumption.
Now, I’m a nobody and I don’t expect what I write to make any difference to anyone. But as primarily a consumer rather than producer of blog posts, it is still completely exasperating when this was what everyone was talking about and the person who started the whole thing (and said that to hold any other position was ‘nonsensical‘) failed to respond at all, except to criticize the seemingly weakest possible example of someone challenging his position.
If anyone else sees it any other way, I’d really like to know because I don’t think I’m being unfair here – unless my expectations about academic civility in the blogosphere are unreasonably onerous.
*Still managed to crank out 943 frustrated words, though…
**I mean this as a
complement compliment (woops!) to Bob
Something I’ve been meaning to talk about for a while is the ‘ideal/non-ideal’ distinction in political theory. A lot of ink has been spilled on exactly what the distinction is or should be, and up front I will say I haven’t read most of it (I really hope to soon get round to reading my former political theory tutor Lea Ypi‘s new book ‘Global Justice and Avant-Garde Political Agency‘, which has a discussion on the topic. Lea, if I screw this post up in some horribly obvious way – I’m sorry!). But the gist (I think) of the distinction is this: when doing ideal theory, we are thinking about what the perfectly just world would look like. When doing non-ideal theory, we are thinking about what justice demands of us in the imperfect world in which we actually live.
One way of thinking about what we are doing in political theory is that we establish what the ideal is (ideal theory), and then figure out how to make the actual world look as much like the ideal world as possible (non-ideal theory). I have a problem with this way of thinking about political philosophy. And it’s essentially the same problem I have with economic theories that are deduced from assumptions that are unrealistic, and whose conclusions are not backed up through some other methodology. The Lipsey-Lancaster Theorem ( or ‘Theory of Second Best‘) in economics states that when one of the optimality conditions for a theory cannot be satisfied, it does not follow that the that other optimality conditions still hold. That is to say, if your proof that x is optimal requires y and z, and y does not hold, it is not necessarily the case that the best alternative to x has z as an optimality condition. When I first read the paper, by Richard Lipsey and Kelvin Lancaster, my mind was blown. But it’s just a particular instance of a very general phenomenon: that when you have a deductive argument following from certain premises, and then weaken one of the premises – all bets are off. It may be the case that a weakened premise can still support a weakened conclusion, but it’s equally possible that nothing follows at all.
I think this relates directly to the ideal/non-ideal theory question, because once you have weakened one of your assumptions about , say, what people are actually like (or especially what people can know, which is almost always ignored) then it just doesn’t follow that moving the world towards something more closely approximating the just world (as derived from ideal premises) is actually what justice demands of us. I take this to be a simple point of logic.
I think this can make the question of advocacy really difficult. If there is anything I have ever learned, it is that often an ‘answer’ to a problem requires a number of distinct elements in order to work. Once you take one of those elements away from me, I have to completely rethink what the right answer is (I find this to be especially true when thinking about financial regulation). If my ‘ideal’ answer would be a world featuring x, y and z, and I can’t have z, it simply doesn’t follow that I should still want x and y. X and y might be a deadly combination on their own! (For example, read x as ‘capital requirements for banks based on risk-weighted assets’, y as ‘having ratings agencies assess the riskiness of assets for regulatory purposes’ and z as ‘competition and competence in the ratings agency business’. We didn’t have z, and it worked out really badly). This means that people can so easily end up talking past each other, because they have different implicit assumptions as to what possibilities are allowed within the particular ‘non-ideal’ rules of the debate.
Of course, too often I use this as an excuse to be lazy about advocacy. I definitely have a tendency to go too far in the sceptical direction, and just throw my hands up in the air and say I have no idea what to do (although when it comes to effectively relieving deprivation, I generally trust GiveWell). But maybe I see difficulty where there is none. If that is the case, I would very much like to know.
Given the balance of the things I write about, I get the feeling that people probably think I’m pretty conservative. I write quite a lot about taxation, the inability of the government to ‘stimulate’ the economy through spending and the fact that states are by nature coercive. And while I have become more conservative in the last 2-3 years, it is at the margin. Now, it just so happens that this is also the stuff that I find really interesting and I just feel like I have more things to say about it. It’s also the case that the standard of argument around these topics is extremely high in the blogs I read, and it’s more fun to get involved in that rather than amateur political philosophy (and Matt Yglesias has pretty much has that angle covered, with appropriate snarkiness).
But I have a whole set of prior beliefs than I haven’t written about nearly as much, but that massively inform my political thinking. For example, I think there are incredibly compelling arguments to the effect that I owe the relative level of my income to factors way beyond my control. Imagine tomorrow everyone woke up and loads more people were good at consulting and not many people at all were good at flipping burgers. Prediction: my salary goes way down, fast food worker’s salaries go way up. My salary is a function of the supply of people offering my skillset, and the demand by other people for those skills. And whatever you think about free will etc., I certainly don’t control other people’s desires or skillsets. I’m but a tiny cog in a massive machine called ‘the labour market’.
That kind of thinking makes me pretty egalitarian. And if total redistribution of income was totally costless, I’d be pretty inclined towards it. But it very much isn’t costless, and I think it would have appalling practical consequences. If human beings were perfectly charitable and unselfish, then it would be fine. But we aren’t. We value ourselves over others. We get greedy when we get power, and massive income redistribution through the government creates a massive concentration of power. We always, always need to be thinking about what will in fact happen if we try and change something.
And this is why I write about the things I do. Because I think people don’t really understand the consequences of policy – and even if they happen to be right, they are almost certainly way too sure about it. The fact that we probably understate the effective tax rate as a percentage of their income for people with savings* doesn’t remotely mean (if correct) that the capital gains tax rate should be zero, or even less than it is now. But it probably does mean that you should have a weaker preference for the tax at the margin. I want you to update your beliefs to reflect new information. Where you end up at depends on where you started (FWIW, I started out thinking they should be taxed the same. I now think it is appropriate that cap gains rates should be lower than income tax rates).
One of the best books I’ve read in the last year is Selfish Reasons to Have More Kids by Bryan Caplan. Whilst the book as a whole is fantastic, it’s the structure of Bryan’s argument that is so wonderful. Bryan presents tons of evidence from twin adoption studies to suggest that genetics has a much larger role to play in the kind of person you turn out to be in the long run than the environment you grew up in**. If true, this means that raising kids need be nowhere near as stressful as most people make it to be. Suppose Bryan is right. If Bryan is right, and your beliefs on the subject of the effect of parenting on children has changed, you should have more kids than you were planning to before hand. Why? Because the cost of having them has fallen (individual parenting decisions aren’t as big a deal as you thought, so you can stop stressing about it). General rule: if the cost of something falls, or the benefits of something increase – you should want more of that thing. This is still consistent with wanting none of that thing, if the costs still outweigh the benefits. But it’s a great example of updating our beliefs to reflect new information (and I highly recommend the book).
I can’t present you with a complete and coherent position on anything. There’s just too many things to know, too many factors to consider and I’m not clever enough. Don’t get too hung up trying to figure out what I think. I’m not too sure myself, a lot of the time! Think at the margin, for yourself. Is this right, is this wrong, how does it affect what I already believe, should I be so sure. It’s not the natural way to think, but it’s the smart way.
* note: that statement depends on the fact that savings were at one point earned, which is obviously not true in the case of inheritance or expropriation
**provided the environment is such that adoption would be legally approved
Can’t remember how (H/T to someone), but recently came across this paper by Milton Friedman from 1966, entitled ‘The Methodology of Positive Economics’. It says pretty much everything I could ever want to say on the subject, and so much more. Here are a few fantastic excerpts:
Viewed as a language, theory has no substantive content; it is a set of tautologies. Its function is to serve as a filing system for organizing empirical material and facilitating our understanding of it; and the criteria by which it is to be judged are those appropriate to a filing system. Are, the categories clearly and precisely defined? Are they exhaustive? Do we know where to file each individual, item, or is there considerable ambiguity? Is the system of headings and subheadings so designed that we can quickly find an item we want, or must we hunt from place to place? Are the items we shall want to consider jointly filed together? Does the filing system avoid elaborate cross-references?
The answers to these questions depend partly on logical, partly on factual, considerations. The canons of formal logic alone can show whether a particular language is complete and consistent, that is, whether propositions in the language are “right” or “wrong.” Factual evidence alone can show whether the categories of the “analytical filing system” have a meaningful empirical counterpart, that is, whether they are useful in analyzing a particular class of concrete problems.
Viewed as a body of substantive hypotheses, theory is to be judged by its predictive power for the class of phenomena which it is intended to “explain.” Only factual evidence can show whether it is “right” or “wrong” or, better, tentatively “accepted” as valid or “rejected.” As I shall argue at greater length below, the only relevant test of the validity of a hypothesis is comparison of its predictions with experience.
And, finally, this
One confusion that has been particularly rife and has done much damage is confusion about the role of “assumptions” in economic analysis. A meaningful scientific hypothesis or theory typically asserts that certain forces are, and other forces are not, important in understanding a particular class of phenomena. It is frequently convenient to present such a hypothesis by stating that the phenomena it is desired to predict behave in the world of observation as if they occurred in a hypothetical and highly simplified world containing only the forces that the hypothesis asserts to be important. In general, there is more than one way to formulate such a description – more than one set of “assumptions” in terms of which the theory can be presented. The choice among such alternative assumptions is made on the grounds of the resulting economy, clarity, and precision in presenting the hypothesis; their capacity to bring indirect evidence to bear on the validity of the hypothesis by suggesting some of its implications that can be readily checked with observation or by bringing out its connection with other hypotheses dealing with related phenomena; and similar considerations.
Such a theory cannot be tested by comparing its “assumptions” directly with “reality.” Indeed, there is no meaningful way in which this can be done. Complete “realism” is clearly unattainable, and the question whether a theory is realistic “enough” can be settled only by seeing whether it yields predictions that are good enough for the purpose in hand or that are better than predictions from alternative theories. Yet the belief that a theory can be tested by the realism of its assumptions independently of the accuracy of its predictions is widespread and the source of much of the perennial criticism of economic theory as unrealistic. Such criticism is largely irrelevant, and, in consequence, most attempts to reform economic theory that it has stimulated have been unsuccessful.
If you had told me three years ago I would be approvingly quoting Milton Friedman, I would have likely laughed out loud. If you haven’t yet, go read the paper. It’s brilliant.