miscellaneous factZ – The online home of Rufus Pollock

From Laslett ‘Phillipe Ariès and “La Famille”‘ p.83 (quoted in Eisenstein, p.131):

The actual reality, the tangible quality of community life in earlier towns or villages … is puzzling … and only too susceptible to sentimentalisation. People seem to want to believe that there was a time when every one belonged to an active, supportive local society, providing a palpable framework for everyday life. But we find that the phenomenon itself and its passing — if that is what, in fact happened– perpetually elude our grasp.

In doing research for the EU Public Domain project (as here and here) we are often handling large datasets, for example one national library’s list of pre-1960 books stretched to over 4 million items. In such a situation, an algorithm’s speed (and space) can really matter. To illustrate, consider our ‘loading’ algorithm — i.e. the algorithm to load MARC records into the DB, which had the following steps:

1. Do a simple load: i.e. for each catalogue entry create a new Item and new Persons for any authors listed
2. “Consolidate” all the duplicate Persons, i.e. a Person who is really the same but for whom we create duplicate DB entries in part 1 (we can do this because MARC cataloguers try to uniquely identify authors based on name + birth date + death date).
3. [Not discussed here] Consolidate “items” to “works” (associate multiple items (i.e. distinct catalogue entries) of, say, a Christmas Carol, to a single “work”)

The first part of this worked great: on a 1 million record load we averaged between 8s and 25s (depending on hardware, DB backend etc) per thousand records with speed fairly constant throughout (so that’s between 2.5 and 7.5h to load the whole lot). Unfortunately, at the consolidate stage we ran into problems: for a 1 million item DB there were several 100 thousand consolidations and we were averaging only 900s per 1000 consolidations! (This also scaled significantly with DB size: a 35k records DB averaged 55s per 1000). This would mean a full run would require several days! Even worse, because of the form of the algorithm (all the consolidation for a given person were done as a batch) we ran into memory issues on big datasets with some machines.

To address this we switched to performing “consolidation” on load, i.e. when creating each Item for a catalogue entry we’d search for existing authors who matched the information we had on that record. Unfortunately this had a huge impact on the load: time grew superlinearly and had already reached 300s per 1000 records at the 100k mark having started at 40 — Figure 1 plots this relationship. By extrapolation, 1M records would take 100 hours plus — almost a week!

At this point we went back to the original approach and tried optimizing the consolidation, first by switching to pure sql and then by adding some indexes on join tables (I’d always thought that foreign keys were auto indexed but it turned out not to be the case!). The first of these changes solved the memory issues, while the second resolved the speed problems providing a speedup of more than 30x (30s per 1000 rather 900s) and reduced the processing time from several days to a few hours.

Many more examples of this kind of issue could be provided. However, this one already serves to illustrate the two main points:

• With large datasets speed really matters
• Even with optimization algorithms can take a substantial time to run

Both of these have a significant impact on the speed, and form, of the development process. First, because one has to spend time optimizing and profiling — which like all experimentation is time-consuming. Second because longer run-times directly impact the rate at which results are obtained and development can proceed — often bugs or improvements only become obvious once one has run on a large dataset, plus any change to an algorithm that alters output requires that it be rerun.

Figure 1: Load time when doing consolidation on load

Last Thursday I attended a talk by Frederick Scherer at the [Judge] entitled: “Deregulatory Roots of the Current Financial Crisis”. Below are some sketchy notes.

Notes

Macro story:

• Huge current account deficit for last 10-15 years
  • Expansionary Fed policy has permitted this to happen while interest rates are low
• Median real income has not risen since the mid-1970s
  • Cheap money mean personal savings have dropped consistently: 1970s ~ 7%, 2000s ~ 1%
• Basically overconsumption

Micro story:

• Back in the old days, banking was very dull — three threes story, “One reason I never worked in the financial industry: it was very dull when I got my MBA in 1958″
• S&L story of 1980s: inflation squeeze + Reagan deregulation
  • FMs: Fannie Mae, Freddie Mac get more prominent
  • [Ed]: main focus here was on pressure for S&L to find better returns without much mention of the thoughtlessness of Reagan deregulatory approach (deposits still insured but S&L can now invest in anything) and the fraud and waste it engendered — see “Big Money Crime: Fraud and Politics in the Savings and Loan Crisis” by Kitty Calavita, Henry N. Pontell, and Robert Tillman
• In 1920s there were $2 billion of securitized mortgages (securitazation before the 1980s!)
• Market vs. bank finance for mortgages: market more than bank by mid-1980s [ed: I think -- graph hard to read]
• To start with: FMs pretty tough when giving mortgages, but with new securitizers and lots of cheap money, standards dropped => moral hazard for issuers [ed: not quite sure why this is moral hazard -- securitizers aren't the ones who should care, it's the buyers who should care]
• Even if issuers don’t care, buyers of securitized mortgages should care and they depended on ratings agencies (Moodys, S&P etc)
• Unfortunately, ratings agencies had serious conflicts of interest as they were paid to do ratings by firms issuing the securities! Result: ratings weren’t done well
• Worse: people ignored systemic risk in the housing market and therefore made far too low assessment of risk of these securities [ed: ignoring systemic risks implies underestimating correlations -- especially for negative changes -- between different mortgage types (geographic, owner-type etc). Interesting here to go back and read the quarterly statement from FM in summer 2008 which claims exactly this underestimate.]
• Banks over-leveraged for the classic reason (it raises your profits if things are good — but you can get wiped out if things are bad)
  • This made banks very profitable: by mid 2000s financial corporations accounted for 30% of all US corporate profits
  • Huge and (unjustified relative to other sectors) wage levels. Fascinating evidence here provide by correlating wage premia to deregulation: fig 6 from Philippson and Reshi shows dramatic association of wage premium (corrected for observable skills) with (de)regulation. Wage premium goes from ~1.6 in 1920s to <1.1 in 1960s and 70s and then back up to 1.6/1.7 in mid 2000s
• Credit default swaps and default insurance: not entirely new but doubled every year from 2001 to the present ($919 billion in 2001 to $62.2 trillion in 2007)
  • Much of the time CDS issued without any holding of the underlying asset
  • There was discussion on regulating CDSes in 1990s (blue-ribbon panel reported in 1998) but due to shenanigans in the house and senate led by Phil Graham (husband of Wendy Graham who was head of Commodity Futures … Board), CDSes were entirely deregulated via act tacked onto Health-Education-Appropriations bill in 2001.

It goes bad:

• Housing bubble breaks in 2007 or even 2006
  • Notices of default starts trending upwards in mid 2006
• [ran out of time]

What is to be done:

• Need simple, clear rules
  • A regulator cannot monitor everything day-to-day
• Outlaw Credit Default Swaps
• Anyone who issues CDOs must “keep skin in the game”
• Leverage ratios. Perhaps? Hard to regulate.
• Deal with too big to fail by making it hard for “giants to form” and breaking up existing over-large conglomerates
• We need to remember history!

Own Comments

This was an excellent presentation though, as was intended, it was more a summary of existing material than a presentation of anything “new”.

Not sure I was convinced by the “remember history” logic. It is always easy to be wise after the event and say “Oh look how similar this all was to 1929″. However, not only is this unconvincing analytically — it is really hard to fit trends in advance with any precision (every business cycle is different), but before the event there are always plenty of people (and lobbyists) arguing that everything is fine and we shouldn’t interfere. Summary: Awareness of history is all very well but it does not provide anything like the precision to support pre-emptive action. As such it is not really clear what “awareness of history” buys us.

More convincing to me (and one could argue this still has some “awareness of history in it) are actions like the following:

1. Worry about incentives in general and the principal-agent problem in particular. Try to ensure long-termism and prevent overly short-term and high-powered contracts (which essentially end up looking like an call option).

   Since incentives can be hard to regulate directly one may need to work via legislation that affects the general structure of the industry (e.g. Glass-Stegall).

   Summary: banking should be a reasonably dull profession with skill-adjusted wage rates similar to other sectors of the economy. If things get too exciting it is an indicator that incentives are out of line and things are likely to go wrong (quite apart from the inefficiency of having all those smart people pricing derivatives rather than doing something else!)

2. Be cautious regarding financial innovation especially where new products are complex. New products have little “track record” on which to base assessments of their benefits and risks and complexity makes this worse.

   In particular, complexity worsens the principal-agent problem for “regulators” both within and outside firms (how can I decide what bonus you deserve if I don’t understand the riskiness and payoff structure of the products you’ve sold?). Valuation of many financial products such as derivatives depend heavily — and subtly — on assumptions regarding the distribution of returns of underlying assets (stocks, bonds etc).

   If it is not clear what innovation — and complexity — are buying us we should steer clear, or at least be very cautious. As Scherer pointed out (in response to a question), there is little evidence that the explosion in variety and complexity of financial products since the 80s has actually done anything to make finance more efficient, e.g. by reducing the cost of capital to firms. Of course, it is very difficult to assess the benefits of innovation in any industry, let alone finance, but the basic point that 1940s through 1970s (dull banking) saw as much “growth” in the real economy as the 1980s-2000s (exciting banking) should make us think twice about how much complexity and innovation we need in financial products.

Finally, and on a more theoretical note, I’d also like to have seen more discussion about exactly why standard backward recursion/rational market logic fails here and what implications do the answers have for markets and their regulation. In particular, one would like to know doesn’t knowledge of a bubbles existence in period T lead to its unwinding (and hence by backward recursion to its unwinding in period T-1, and then T-2 etc until the bubble never existed). There are various answers to this in the literature based on things like herding, presence of noise investors, uncertainty about termination, but it would be good to have a summary, especially as regards welfare implications (are bubbles good?), and what policy interventions different theories prescribe.

I’m posting up an essay on “Discounting and Self-Control” (pdf). The essay, which I haven’t really touched for over a year, is still in its early stages but having lacked the time to do much on it over the last year, and going on the motto of “release early, release often”, I’m posting it up as a form of alpha version.

… then must you speak
Of one that loved not wisely, but too well;
Of one not easily jealous, but, being wrought,
Perplex’d in the extreme; of one whose hand,
Like the base Judean, threw a pearl away
Richer than all his tribe; …

– Othello, The Moor of Venice

Abstract

An agent’s intertemporal choices depend on a variety of factors, most prominently, their valuation of future payoffs as encapsulated in a discount function. However, it is also clear that factors such as self-control may also play an important role, and given the similarity of impact, a confouding one. We explore the literature on this issue as well as examining what occurs when those with higher time-preference (whether arising from discounting or self-control) also enjoy their consumption more.

Introduction

The exercise of will, especially in the form of self-control, has long been recognized as central to human existence, experience, and morality. Over the last few decades there has been increasing interest in the issue from a scientific perspective. At the same time, it has also long been appreciated that humans (and other animals) make trade-offs between the present and the future — as well as between different points in the future, and that events taking place closer to the present are given greater weight than those which are more distant. Traditionally, at least in economics, this type of behaviour has been subsumed under the heading of discounting.

Both of these factors, self-control and discounting, affect behaviour, and choices, in relation to outcomes which do not (all) take place in the present. However they are distinct. Specifically, consider a very simple case of two outcomes A and B where B occurs after A (for example, A might be one ice cream today and B an ice cream and a doughnut tomorrow). Self-control issues arise where one prefers B over A but is unable to execute on this preference and therefore actually takes (’chooses’) A. By contrast, in the discounting case A is actually preferred over B and therefore is chosen (freely) by the decision maker.

It would seem important to keep these two aspects of decision making clearly separated. While lack of ’self-control’ is usually seen as disadvantageous and a reason for adopting various ‘commitment strategies’ — for example, by opting to remove various items from the choice set (having no cigarettes in the house) — the simple preference for the present over the future incorporated in the discounting model would seem to generate no such difficulties.

However, empirically it may prove rather difficult to do so. As shown by the simple example above the same observed ‘choice’ for A (one ice cream today) over B (ice cream plus doughnut tomorrow) can be the result of two very different processes. Thus if we only observe choices, and not the underlying preferences and/or the process by which the choice is arrived at, it may be impossible to distinguish the two.

It is perhaps for this reason that these distinct aspects are sometimes conflated. Consider, for example, Mischel et al 1989 which is entitled “Delay of Gratification in Children” and summarizes much of Mischel of pioneering work on this area. Mischel’s approach is clearly more oriented along the self-control aspect, and this is borne out in the types of experiments conducted (more on this below). Nevertheless they state (p.934) “The obtained concurrent associations [between treatments and delay] are extensive, indicating that such preferences reflect a meaningful dimension of individual differences, and point to some of the many determinants and correlates of decisions to delay (18).” Here the orientation towards self-control has become a general “decision to delay” and this is borne out by the associated footnote (18) which references related literature in other disciplines and is worth quoting in its entirety:

[... see full essay for more]

Lots of people have been up in arms about a letter sent out by Ordnance Survey about the “Use of Google Maps for display and promotion purposes”. With titles like “Are the Show Us A Better Way winners safe from Ordnance Survey?” (Guardian), “Home Secretary’s crime maps not allowed say Ordnance Survey” (localgov.co.uk) or “The mapping mess – Google v OS” (bbc.co.uk) these seemed to indicate some particularly unreasonable behaviour by OS.

However, after actually reading the original OS letter I’m far from convinced. In essence OS say:

• If you have created the data yourself you can do whatever you like with it including plotting it on Google maps.
• However, if you have derived the data from an OS map then you can’t. You can’t because a) as derived data OS have rights in it b) plotting it on a Google map according to Google’s to T&C gives Google a perpetual, royalty-free license to that data. Since, unsurprisingly (and not unreasonably) OS don’t want to give Google such a license they don’t want you plotting it on a Google map.
• What the implications of this are then depend heavily on:
  1. Whether Google will change their licensing conditions (at least for their free service)
  2. What derived data is

Much of the discussion centred on the last of these items: what is derived data? OS state:

Simply put, Ordnance Survey derived data is any data created using Ordnance Survey base data. For example, if you capture a polygon or a point or any other feature using any Ordnance Survey data, either in its data form or as a background context to the polygon/point/other feature capture, this would constitute derived data.

It should also be borne in mind that data from other suppliers may be based on Ordnance Survey material, and thus the above considerations may still apply. We therefore recommend that you verify whether any third-party mapping you use may have been created in some way from Ordnance Survey data before displaying it on Google Maps.

NOTE: Again, the answer to this question is based on our understanding of which of Google’s standard terms and conditions we believe would apply. In the event that Google is prepared to offer you terms and conditions which do not involve you purporting to grant Google a licence of Ordnance Survey base or derived data, we would have no objection to your hosting such data on top of Google Maps in this scenario.

My understanding of this is that if you extract the geodata from an OS map (i.e. polygon, points, features) by some extraction method (such as tracing) then that’s derived data and OS can control what you do. This is pretty standard: if I copy text from a book by typing it out longhand I’m still infringing copyright.

However, this does not mean if I’m using OS maps as a base-layer and, for example, by clicking at some particular point I generate a lat-long (say to indicate where I live, or where a crime happened) then that lat-long is ‘derived’ data.

Now, of course, this could be a fine line: if I happened to click on a bunch of points, say to indicate a walk I went on, and these also showed the route of road there could be debate as to whether I’m infringing the OS rights in the feature or not.

Nevertheless, the basic principle (as I understand it) is clear: geodata created when using OS tools and maps is always yours unless it is directly replicating the underlying OS data. If this interpretation is correct then this whole debate is a bit of a storm in a teacup and projects such as crime-mapping or providing a loofinder aren’t at any risk from OS’s licensing terms.

It would be interesting to chart over time the progress of open-source, standards compliant, Mozilla-type web browsers (e.g. Firefox) versus Microsoft’s Internet Explorer. As is often the case in other areas, it is not easy to get good (open) data over a reasonable time period. The graph below shows browser market share as measured by the browser usage of visitors to the W3Schools website (data source on Open Economics plus the code to extract original data into this usable form).

Browser Market Share (NB: Firefox was released in Nov 2004 but not listed separately by W3Schools until 2005)

Given the source, and therefore the bias towards more technically savvy users, these figures probably overstate Firefox’s market share somewhat, though the overall trend is probably largely correct. What we see is a steady and continuing increase in Mozilla (Firefox) market share ever since Firefox’s launch in Autumn 2004 and a concomitant decline in market share of IE (the little dip for Firefox at the end appears to be directly attributable to the launch of Google’s chrome). What is particularly interesting is that, at least for W3Schools users, we are almost at the point where there are as many people using Firefox as IE. This is significant for several reasons.

First, because of its level of usage it will no longer be possible for websites to only ‘work in IE’ but instead will always have to work in Firefox as well. This is both good for Firefox and for the standards-compliant browsers more generally (while, of course, Firefox itself is not perfectly standards compliant it has traditionally been much better than IE).

Second, it is an (unusual) example of a case where dominance has not been maintained. Generally a firm with established dominance in a given area is able to maintain — witness the robustness of Microsoft’s established dominance in other areas. By contrast, in this market, as the graph shows, Firefox has almost drawn level with IE and may soon surpass it if the trend of the last few years continues.

The human problem of ’scarce resources and unlimited wants’ is oft-posited as a primary motivation for studying economics. As this phrase makes clear, ‘wants’ (’preferences’ to use the more usual terminology) are a central part of what we study, and the existence, and stability, of those ‘wants/preferences’ therefore merit serious consideration.[^1]

[^1]: It is interesting how the term ‘preference’ is studiously neutral, and almost anodyne in comparison with a term such as ‘want’, ‘desire’ or even ‘need’, each of which is a potential synonym. One might imagine, and this is simply conjecture, that the term was intentionally adopted in order to remove any overtone of judgement. After all, across most culture and over much of human history, the formation and satisfaction of ‘preferences’ has been a process laden with ethical, and religious, significance.

Few of us have difficulty accepting the fundamental nature of our desire for food and shelter. However, many of us might have greater difficulties assigning the same fundamentality to the desire for a particular brand of designer perfume or a digital music player. In fact, it is unclear to what extent one can want what one has never known (or conceived of), and thus, while it is not difficult to imagine any human desiring food and shelter — especially when they are absent, it is hard to imagine a stone-age nomad, say, even being able to conceive of designer perfume or iPods (let alone feel their lack).

It is also telling that so many of the consumer goods, especially those away from the necessity end of the spectrum, appear to require active promotion to the public. Of course it is true, as economists are particularly fond of pointing out, that advertising has an informational component — simply letting you know about the existence and attributes of products. However, it is also hard to deny that advertising also has a substantial ‘persuasive’ component, operating either to create preferences or alter existing ones.

If so this has important implications. In particular, it strongly suggests that our wants aren’t simply given but are, at least to some extent, formed by our experience and choices.[^2] This raises some deep and important questions for economists to answer — questions with a major bearing on the state and direction of many modern societies. It also has some direct connections with one of the oldest, and most philosophical, of the world’s religious traditions: Buddhism. Central to Buddhist teaching are the Four Noble Truths. Succintly put these are, in order:[^3]

[^2]: In economics jargon: preference are endogenous (i.e. determined within the system) rather than exogenous (fixed externally — e.g. by ‘nature’). The study of endogenous preferences is certainly not new. See for example the review of Bowles (1998) or the early incorporation of changeable preferences into the ‘traditional’ framework by Becker and Stigler (1977).

[^3]: These translations of the Dhammacakkappavattana Sutta are taken from http://www.accesstoinsight.org/tipitaka/sn/sn56/sn56.011.piya.html

1. (Dukka — The Nature of Suffering) ”Birth is suffering, aging is suffering, sickness is suffering, death is suffering, association with the unpleasant is suffering, dissociation from the pleasant is suffering, not to receive what one desires is suffering — in brief the five aggregates subject to grasping are suffering.”
2. (Samudaya — The Origin of Suffering) ”It is this craving (thirst) which produces re-becoming (rebirth) accompanied by passionate greed, and finding fresh delight now here, and now there, namely craving for sense pleasure, craving for existence and craving for non-existence (self-annihilation).”
3. (Nirodha — The Cessation of Suffering) ”It is the complete cessation of that very craving, giving it up, relinquishing it, liberating oneself from it, and detaching oneself from it.”
4. (Marga — The Path to Cessation of Suffering) ”It is the Noble Eightfold Path, and nothing else, namely: right understanding, right thought, right speech, right action, right livelihood, right effort, right mindfulness and right concentration.”

Why is this teaching relevant here? First, observe a commonality: both economics and Buddhism takes unsatisfied ‘wants’ (or ‘cravings’) as a source of unhappiness. But how do go about solving this problem? Here economics and Buddhism part ways, and rather dramatically, with the Four Noble Truths presenting a path to the achievement of well-being which is almost diametrically opposite to that advocated by economics.

Specifically, the ‘economics’ approach, is based on taking preferences as given and focusing on generating the goods to satisfy them. By contrast, Buddhism sees ‘wants’ as ultimately unsatisfiable, and instead proposes that the way to well-being is not to satisfy them but to relinquish them — while some ‘cravings’ can be temporarily satisfied more will always be generated, moreover there some fundamental desires, such as the wish not to die, cannot be addressed in the material world.

Put starkly: economic thought directs our energy efforts to satisfying our wants taking them as given while Buddhism directs those self-same energies to altering our wants, and views most attempts to satisfy them by obtaining ever more ‘things’ as inevitably doomed to failure — in fact, actively counter-productive as more ‘wants’ are generated by the very process of satisfaction.

Suddenly all talk is of financial and ‘economic’ crisis. Being an economist I am repeatedly being asked for my views on this recent turn of events. I’m not an expert in this area so I’m rather hesitant to give an opinion: my instinctive response is to point out that if I really knew what was going to happen next a) I’d be (or about to be) a very rich man b) I probably wouldn’t be telling them.

My more measured response would be to emphasize a) some basic, crude, but useful ‘facts’ b) the value of getting hard data.

A. We still don’t understand business cycles very well but we do think it very likely that credit availability (or its lack) play a significant part. Severe restrictions on credit associated with financial panic have figured prominently in past downturns. Thus, given the severity of the current credit situation, it is likely that the ‘real’ effects on employment and output will be significant.

B. Financial firms in a financial crisis can fail for two reasons:

1. Illiquidity: i.e. insufficient funds to pay immediate liabilities. This is the classic bank run.
2. Genuine insolvency: assets (even evaluated at a reasonable horizon) are insufficient to cover liabilities.

One of the major purposes of banks is to solve ‘liquidity’ mismatches: I want to lend on call but you want to borrow for 20 years. Because of this a bank even when ’sound’ can if there is a sufficiently large withdrawal of funds in a short enough time period (and borrowed funds cannot be found). This is the classic bank run — though it can also apply to any financial institution which has a potential liquidity mismatch (i.e. creditors have money on call but money to debtors is tied up for a longer period). Clearly, public policy will usually differ across the two cases: whereas public support for a sound institution facing a run seems common sense, ‘propping up’ a bankrupt one seems less sensible.

One of the major difficulties, for a casual observer of the current crisis such as myself, is distinguishing exactly which institutions fall into which category. Was Lehmann Brothers bankrupt of just facing a bank-run, ditto for AIG and Fannie Mae? Moreover, the very fact that uncertainty exists may result in such a freezing of credit — due to fears over counter-party defaults — as to cause failures of Type I as well as Type II institutions (as well as substantial collateral damage to all borrowers — i.e. most businesses).

Also difficult is that fact that Type I can shade into Type II because the forced sale of assets to meet ‘margin calls’ (be these payments to bank depositors or any other creditor) tends to depress asset prices. In the worst-case we have the classic ‘fire-sale’ in which the urgent need for cash leads to forced sales at massively reduced prices. Since this depresses the price not just of the firm selling but for all other firms holding that asset this can initiate a continuing downward spiral of sales (though, of course, we need to explain why there aren’t enough ‘rational buyers’ on the other side to drive prices back to equilibrium — the simple answer being that such buyers are facing the same tightening of credit as everyone else).

In the present case things seem to have been made worse since many (some?) of the assets held by banks were in the form of illiquid and opaque CDOs and CMOs (collateralized debt/mortgage obligations). Since these are rarely traded, pricing them (marking them to market) is not always easy. In times of ‘panic’, mass-selling driven down their price massively, perhaps far below any likely future price level, but forcing all holders (including non-sellers) to mark down their value. Normally, the link between price today and price tomorrow should be enforced via arbitrage but as already noted that link can disappear in the crisis.

At the same time, it is also likely that such securities were overpriced (or under-risked) previously. In particular, many of the mortgage based securities appear to have been overpriced relative to their risk, due to over-optimistic predictions for the future price path of housing, and (in the US), extrapolations of historical geographic price correlations which turned out to be too low (house prices, since these closely relate to the long-term path interest rates and economic growth are especially sensitive to revisions in beliefs about these factors). Similarly, as the possibility of a global recession has turned into seeming certainty, there has been a substantial re-evaluation of expectations regarding business earnings, at least over the next few years which has fed through in stock markets.

Though it is likely, to judge from past experience, that markets have over-reacted, it is not easy to say by how much: at least some of the change is due to genuine changes in expectations about the future paths of real variables and not simply the self-fulfilling dynamic of a crash. Returning to our previous points, it is exactly this uncertainty that helps feed the crisis, and simultaneously the reason that hard data is so important. Exactly how much of what assets are given institutions holding? How far have they already been marked down? What were the implied earnings level of prices a year and a half ago compared with today? How far do house prices in the US have to fall before X goes insolvent? (If the answer is 20% we should be worried if it is 50% we probably should not).

Of course, financial institutions — and others — are loath to disclose this kind of information for precisely the the reason that it may reveal their precarious position. But perhaps the time has come when we need this information out in th open.

This January I met Alan Kirman at the Robinson Workshop on Rationality and Emotions. Over lunch we had a brief discussion about the difficulties of modern macroeconomics. I was therefore intrigued to see a new paper of his (co-authored with Peter Howitt, David Colander, Axel Leijonhufvud and Perry Mehrling) entitled Beyond DSGE Models: Towards an Empirically-Based Macroeconomics which was presented in January at the AEA conference (and looks like it will be appearing in the AER ‘Papers and Proceedings’).

The paper has much to say about the current state of macro, in particular the serious problems with DSGE (dynamic stochastic general equilibrium models) and where we should go from here. As the abstract puts it:

This paper argues that macro models should be as simple as possible, but not more so. Existing models are “more so” by far. It is time for the science of macro to step beyond representative agent, DSGE models and focus more on alternative heterogeneous agent macro models that take agent interaction, complexity, coordination problems and endogenous learning seriously. It further argues that as analytic work on these scientific models continues, policy-relevant models should be more empirically based; policy researchers should not approach the data with theoretical blinders on; instead, they should follow an engineering approach to policy analysis and let the data guide their choice of the relevant theory to apply.

It is worth quoting at some length from the paper in order to bring out the full ramifications of the story the authors tell:

Keynesianism Goes Wrong

With the development of macro econometric models in the 1950s, many of the Keynesian models were presented as having formal underpinnings of microeconomic theory and thus as providing a formal model of the macro economy. Specifically, IS/LM type models were too often presented as being “scientific” in this sense, rather than as the ad hoc engineering models that they were. Selective micro foundations were integrated into sectors of the models which give them the illusory appearance of being based on the axiomatic approach of General Equilibrium theory. This led to the economics of Keynes becoming separated from Keynesian economics.

The Reaction and a New Dawn (Rational Expectations and Neoclassical GE Models)

The exaggerated claims for the macro models of the 1960s led to a justifiable reaction by macroeconomists wanting to “do the science of macro right”, which meant bringing it up to the standards of rigor imposed by the General Equilibrium tradition. Thus, in the 1970s the formal modeling of macro in this spirit began, including work on the micro foundations of macroeconomics, construction of an explicit New Classical macroeconomic model, and the rational expectations approach. All of this work rightfully challenged the rigor of the previous work. The aim was to build a general equilibrium model of the macro economy based on explicit and fully formulated micro foundations.

But ‘Technical’ Difficulties Intervene

Given the difficulties inherent in such an approach, researchers started with a simple analytically tractable macro model which they hoped would be a stepping stone toward a more sensible macro model grounded in microfoundations. The problem is that the simple model was not susceptible to generalization, so the profession languished on the first step; and rational expectations representative agent models mysteriously became the only allowable modeling method. Moreover, such models were directly applied to policy even though they had little or no relevance. … [emphasis added]

But There Was a Reason For This: Other Stuff is Hard

The reason researchers clung to the rational expectations representative agent models for so long is not that they did not recognize their problems, but because of the analytical difficulties involved in moving beyond these models. Dropping the standard assumptions about agent rationality would complicate the already complicated models and abandoning the ad hoc representative agent assumption would leave them face to face with the difficulties raised by Sonnenschein, Mantel and Debreu. While the standard DSGE representative models may look daunting, it is the mathematical sophistication of the analysis and not the models themselves which are difficult. Conceptually, their technical difficulty pales in comparison to models with more realistic specifications: heterogeneous agents, statistical dynamics, multiple equilibria (or no equilibria), and endogenous learning. Yet, it is precisely such models that are needed if we are to start to capture the relevant intricacies of the macro economy.

Building more realistic models along these lines involves enormous work with little immediate payoff; one must either move beyond the extremely restrictive class of economic models to far more complicated analytic macro models, or one must replace the analytic modeling approach with virtual modeling. Happily, both changes are occurring; researchers are beginning to move on to models that attempt to deal with heterogeneous interacting agents, potential emergent macro properties, and behaviorally more varied and more realistic opportunistic agents. The papers in this session describe some of these new approaches. [emphasis added]

Some Closing Comments of My Own

So there you go: plenty of tough challenges and a big dose of humility. To some extent here it seems thing run on 30-40 years cycles: Keynesianism from 1945-1975, Rational Expectations DSGE from 1975-2005 and now we’re into the era of complexity and ‘loose’ tools with emphasis on empirics and heuristics rather than formal models. Whether this new approach will deliver more than the old is yet to be seen. After all, one reason that there are so many physicists getting interested in Economics and Finance is that the going is so hard in, e.g., condensed matter physics (superconductivity anyone …). If the economy really is so complex will we ever do any better at the macro scale than we do for the weather and if so will it not rely on some conceptual breakthrough rather than just doing using more hard-core dynamical systems theory and running more agent-based simulations?

That said, as the authors argue, the ’simple’ route isn’t working and the hardness of the path is no reason not to attempt it — an argument in many ways directly inverse to the traditional ‘drunkard-and-the-lamp’ approach in which we restrict our models, often beyond the point in which they remain relevant, in order to maintain analytical tractability. Thus, though cautious regarding what more ‘complexity-oriented’ methods can deliver, I am in wholehearted agreement with the authors that they justify much greater exploration.

Originally status would have developed from some kind of of stimulus-response setup:

    Beating Competitor
            |
            V
      Higher Status
            |
            V            
Better Access to 'Resources'
  (e.g. Food and Partners)
            |
            V            
  Higher Survival Rate /
    More Progeny etc
            |
            V            
  Development of Reward System(s)
   for these outcomes (Food etc)   
            |
            |  (short-circuiting
            |   as with e.g. sex)
            |
            V         
Development of Reward Systems
 for Success in Competition 
     (Higher Status)

So status now has two components:

1. Increase in status improves access to ‘basic’ goods we derive direct ‘utility’ from (food etc)
2. Increase in status provides direct ‘utility’ independent of any impact upon access ‘basic’ goods.

What about respect? It could be argued that respect is a ‘basic’ good directly equivalent to type (ii) status. However I’m not really convinced of this for two reasons. First, ‘respect’ is fundamentally different from ‘normal’ goods in that one can select what you respect (and whose respect you care about). Second, and more importantly, as just outlined above, the desire for ‘respect’ or ’status’ seems to me a ’secondary’ desire, which has come about via a short-circuiting of our basic reward systems for ‘primary/basic’ goods.

Leaving this aside, the crucial point is that type (ii) status results in a pure zero-sum game. Thus, reducing competition for it (perhaps by increasing compassion) might move us to a (more) positive sum situation. Furthermore, the clear distinction between type (i) and type (ii) allow us to separate out ‘competing to survive’ (which might be essential) and ‘competing (just) to win’. This seems an important distinction to make. After all, we can all accept that, in a whole set of situations, successfully competing may be crucial to obtaining the basic resources needed to survive. However as we get wealthier it would seem that this first aspect diminishes in importance and the second (less healthy) aspect of status looms ever larger.