Although more and more readers are coming around to our point of view on how much bank IT and payment systems issues need to be seen as a first-order issue to be addressed in considering a Grexit, we still encounter a considerable minority of readers and other commentators who find it upsetting to be told that information infrastructure, and even seemingly mundane issues like designing, printing, and distributing currency (as in refitting ATMs to handle it and the logistics of getting the cash securely to ATMs and restocking them) are not trivial issues.
I wonder if this resistance is a symptom of the fragmentation and specialization of the overwhelming majority of work. In the old primitive days of greater self-sufficiency, people would have to worry on a routine basis about how things could screw up and would need to plan: making sure you kept your tools in good repair, knowing how to store foods to get through the winter, knowing the time it took to get to the nearest town for supplies and other essentials (say medical care). By contrast, most Americans who have adequate incomes can shop for food daily if they want to.
The reason we are pressing on with this issue, even though the question of a Grexit is now in abeyance, is that the creaky state of bank IT, and the systemic risk it poses, is only going to get worse. We’ve been gratified to see a lot of readers generously provide a tremendous amount of detailed insider commentary. We plan to continue to focus on this issue and hope the professionals who can give us their war stories and insights remain engaged.
I find it puzzling to encounter what often looks like knee-jerk resistance when readers with deep expertise in bank information technology explain why the lack of documentation of virtually all coding, the large number of different languages used, and the interdependencies among systems means that changing currencies (really, running two currencies in parallel when before you had only one) is a daunting problem for a single bank, and becomes vastly more complicated when you look at issues across the banking/payment system. As reader JustAnObserver wrote:
I’m in the IT biz, h/w not s/w but still …, Louis & Clive (+ other commentators) have pretty much nailed the difficulties of modifying, testing, and then deploying new code into an exiting, running, mission-critical system stuffed fully of dusty deck legacy. They are legion even before you get to what might be called the “user interface” aspects – in this case POSs, ATMs, PIN delivery, check clearing etc. Its not called spaghetti code for nothing (*)
However what has struck me is that there’s an assumption here that seems to be going without question: All the s/w that has to be changed is still available in source code form – be it COBOL, PL/1, Assembler, Fortran or whatever – and can be recompiled. Maybe in amongst the decades old stuff are binary blobs of machine code whose source has long since been lost and whose interface is barely defined/understood, if at all.
On a different note. All this long thread on the IT aspects of Grexit does in my mind is to, once again, raise the question I asked myself a long time ago and to which I have never had any kind of satisfactory answer. Why are those on the left (however strongly/weakly you may define that term) so utterly and pig-headedly resistant to the idea that if you are going to (persuade people to allow you to) change things then you might at least *try* to make the changes work. Here. In the Real. With all its complexities, interconnections, uncertainties.
What exactly is wrong with the left showing a little competence? For a change.
In Louis Proyect’s latest post on the difficulties of re-introducing the drachma, another reader explained why changes that seemed like they ought to be simple in a big bank IT environment were typically anything but. And notice how he is talking about changing foreign exchange systems, precisely the same type that would need to be changed to incorporate a new drachma. From Unknown Gnome:
I’ve been reading this series with interest and thought I could offer some insights on the difficulties of a transition to a new Greek Drachma based on my background and experience.
By way of my background – I have worked for over 10 years in the IT department supporting the foreign exchange front office arm of a bulge bracket bank known for the size of its foreign exchange business.
I joined initially as a programmer, started working as team lead and am now employed by the same bank as a delivery manager of a largely outsourced and offshored IT department.
On the basis of working for my employer I probably have a fairly good insight in to the difficulties of making IT changes to introduce a new currency to a large bank and most banks of our size are likely to face the similar of problems.
So how difficult is it?
I can probably best illustrate this on the basis of a recent change request.
Our foreign exchange sales and trading arm wanted to increase the precision of certain foreign exchange instruments we were quoting – we were supposed to make a global change so that an instrument – one currency quoted vs. another – could be offered to clients at a higher precision.
The request that came in was to increase the precision by one digit – so we could quote rates with 5 digits of type 1.23456 as opposed to 1.2345 we offered previously – this was requested for 200 different currency pairs.
How difficult could this be? We decided to start by changing the precision of just one currency pair in a test environment.
The problem was simplified by the fact that the change wasn’t going to affect the operations/back office components in a big way. After all cash flows and not rates were settled and no changes were going to be required to the risk management or middle office components.
Nevertheless introducing and successfully testing 1 digit precision change for a single currency pair in to a software test environment took over 4 months and changes to 50 separate software applications.
“Why was this so hard anyway?” – a senior MD wanted to know – “couldn’t we IT people keep the instrument definitions in a single central place?”
All I could offer were excuses that sounded lame – even to me – the applications hadn’t been developed with this in mind.
Not surprising really given that our trade capture and price distribution tools – written in Java are 5-10 years old. The risk management and middle office applications are 15-20 years old and are written in C++. The settlement systems are 30+ years old are written in Cobol and run on a mainframe.
Each component had its own and often undocumented source of instrument data.
What does this tell us about the effort required to introduce a new Greek Drachma?
This would require a lot more effort for a number of reasons.
Firstly it isn’t possible to simply revert the original ISO currency code (GDR).
An example will illustrate this – let’s assume a bank had issued a 30 year loan in the original Drachma to a financially sound commercial counterparty in 1998 to mature in 2028. This loan would have been converted at the agreed fixed rate to the Euro in 1999.
Both counterparties should expect to settle in 2028 in Euros at the rate fixed in 1999 – not a new rate for the new Drachma and any attempt to re-activate the original instrument with a different exchange rate would cause problems both in the banks spot and interest rate risk management applications and department as well as the operations department at settlement date.
Lots of applications would be affected.
The front office applications responsible for price distribution and trade capture would need to recognize and price the new Drachma instrument vs other currencies – even in a managed exchange rate environment with the rate fixed by the new central bank of Greece.
Risk management, finance and the global ledger would need to have their risk decomposition and book assignment rules upgraded to ensure the correct aggregate books contained the correct amounts.
Several sources of instrument static data would need to be upgraded.
The settlement systems would need to be upgraded to allow them to confirm trades and settle transactions in the new Drachma.
Central pieces of essential inter-bank plumbing run among others by Swift and Reuters would need to be upgraded as well. It should be noted that these pieces are likely to be required even for the lesser goal of enabling financial transactions between Greek banks within Greece in the new Drachma.
It’s worth remembering here that when the Euro was introduced commercial banks and other financial intermediaries were given multiple years to do the work and a strong commercial incentive of the common market in a single currency.
A potential disorderly exit from the Euro by Greece would leave many banks with little time to implement the changes and – short of coercion by the EU – little commercial incentive given Greece’s much smaller economy.
Without the global IT plumbing modern cross border financial transactions in the new Drachma would be all but impossible for Greece.
In case of a Grexit their best bet in the interest of keeping any form of financial exchange going would probably be to unofficially adopt the Euro or US Dollar ad interim – the way Ecuador did in 2000.
Yves here. The wee problem with “unofficially adopting” the Euro or dollar is it is completely at odds with the objectives for reintroducing the drachma: first, for Greece to be able to attain monetary sovereignity, and related to that, to have the drachma become accepted for international/cross border transactions so that Greek exporters would gain the benefits of having a cheaper currency. You can only do so much having people use foreign currencies in physical form in Greece and having locals all run de facto foreign exchange businesses by making their own conversions to drachma in the course of conducting business. Our expert Clive explained at some length, why converting electronic point of sale systems is not trivial and why that matters. For instance, credit cards are widely used by tourists on vacation, and for good reason:
To those who might suggest that Greece could revert to a cash only society for a time until the card payments system was modified it is worth noting that for some industries – notably lodging, rail travel, car hire (all of which are important sectors in Greece because they are part of the tourist industry) – have a very high percentage of settlement via card payments. This is because customers do not carry sufficient cash to cover these larger bills or even, for vacations especially, do not actually have the cash available and need to tap a line of credit. And for car rental, for example, some sectors have evolved to become entirely dependent on card payments. It is very difficult to hire a car without a card being used to pre-authorise the likely bill and to allow the car rental company to get an irrevocable deposit should you damage the car. If you don’t have a card to use, you need a huge cash deposit, often $1000+.
And cash societies are very susceptible to regressing to a black market economy with the predicable implications for a government’s ability to secure tax receipts.
A remark by Joe Firestone in a recent post seemed to encapsulate the core bone of contention, and Clive graciously responded via e-mail. First, from Firestone:
One thing I do know is that “solutions” to the IT problem of Grexit that take up to three years to implement are not acceptable solutions to the IT problem, since Greece may have to Grexit in the short run without very much of a grace period.”
It all became clear to me that, in relation to a Grexit and IT, Joe has no intention at all of engaging with the facts let alone any theory. He only has a hammer (that the Greek financial systems’ IT must be “fixed” in a few weeks or at most a few months to handle a Grexit) so every time he peruses the matter he goes looking for things that he hopes are nails (“solutions” that can be implemented in a few weeks or months).
If anyone like me dares to suggest that there are no nails, or anyway not the sorts of nails he wants to be provided, I get accused of not trying hard enough to find them. He himself claims that he might have found objects which could be nails but he doesn’t know enough about nails to identify them correctly. Joe then invites (or professes to invite) opinions on whether what he has found are indeed nails. But again, if he gets back contratrian voices saying, no, those aren’t nails, they’re patent medicine in fancy bottles, we get told that we just don’t recognise a good nail when we have it put in front of our noses.
Unfortunately this type of thinking is not uncommon. On a pretty weekly basis I get someone coming to me to tell me that s\he has got this project they want doing, it is all terribly simple, they are expecting that I’ll have it completed in a month or so and it will cost about 5 pence. I send them away telling them to come back when they’ve gotten real. But if they persist, then the problem is self-resolving. Either they’ll ask around some more and get the same asnwer I’ve already given. Or else they’ll find someone who will try to do whatever impossible thing is being demanded – and then it is apparent withing a month or so at the most what is reality and what is delusion.
Unfortunately in a blog and comments it’s not so easily settled. Joe is completely convinced that with all this sh*t piled up, there must be a pony here somewhere. He then invites (demands?), not unlike Hillary and BLM, that others come on in and get shovelling. If you try to tell him it’s all nothing but a pile of crap, you just get given a bigger shovel.
One reason we’ve kept on about this issue is that the stakes for Greece are much higher than just wasting time and money on the IT equivalent of quack cures. Greece is not self-sufficient in food, energy, or pharmaceuticals. Greece also has a tourism sector that is 18% of GDP. Greece got a mild taste of what impaired access to a payments system amounted to in its two-week bank holiday. Tourist bookings did indeed collapse, food shortages were starting to occur, importers struggled mightily to survive and many continued to suffer significant damage even after the banks began to get liquidity. The fact that the Greeks themselves reject the idea of a Grexit despite the horrific costs of continued austerity suggests they have a better grasp of the cost and benefits than well-intended foreigners do.