U.S. incumbent local-exchange telephone companies’ average cost per loop has risen equivalent to 1.9% per year from 1988 to 2007. That’s a 0.4% reduction per year in real (inflation-adjusted) terms.[1] In contrast, computing costs and storage costs have probably been falling more than 40% per year in real terms over the past two decades. Thus telephone companies’ cost growth has been roughly an order of magnitude worse than that computing and storage cost trends.
Narrower telephone company cost categories have shown only slightly better cost trends. Over the past two decades, telephone company network operations expenses and corporate operating expenses have fallen 2.6% and 1.5% per year in real terms, respectively.[2] Network operations and corporate operations are general administrative functions for which information technology typically is crucial. But large cost-performance improvements in information technology are not apparent in these telephone company cost accounts.
Reported telephone company network structure shows little indication of technological change. The number of central office switches per telephone company state-specific service area (study areas or COSAs) has decreased little since the mid-1990s. The number of switched access telephone lines per switch has increased little since the mid-1990s. These indicators of network structure show no evidence of increasingly powerful network switchers / routers.
Switching technology in fact has gotten so powerful that it’s not described in relation to telephone service. Cisco’s recently announced router, the CRS-3, has capacity equivalent to routing simultaneous video calls for every man, woman and child in China. All U.S. telephone calls wouldn’t be noticed within this one router’s switching capacity.
Telephone companies’ network operations costs far exceed this router’s cost. According to Cisco, the price of the CRS-3 starts at $90,000. U.S. telephone companies’ annual reported network operating expenses are about 73 thousand times as large as that price. Costs other than than technological costs of providing telephone service account for most of telephone companies’ costs.
The particular organization of businesses and markets can generate large costs. For example, in the mid-1990s, annual advertising and marketing expenses for U.S. long-distance telephone services were comparable to the total capital budget for building a new national, high-capacity network.
With a different organization of communications companies and communications markets, telephone service costs could look like email service costs.
* * * * *
Data: U.S. telcos non-traffic-sensitive costs and network structure, 1988-2007 (Excel version)
Notes:
[1] These figures are for non-traffic-sensitive costs per loop as defined under regulations for determining high-cost loop subsidies. See 47 CFR 36, Subpart F. Incumbent local exchange telephone companies (ILECs) are local exchange telephone companies that provided service prior to 1996.
[2] Network operations expenses are the costs reported in account 6530 as defined in 47 CFR 32.6530. Corporate operating expenses are costs reported in accounts 6710 and 6720, as defined in 47 CFR 32.6710 and 47 CRF 32.6720. Again, these figures are for ILECs only.
And your point is?
If you think it is so cheap to offer telephone service, why aren’t investors pouring money into start-ups to do it?
dear douglas,
you have a very active mind. great work, see you 2011 at TPRC if you attend.
Best,
Aine (marie Patricia) Ni Shuilleabhain
will write more soon.
Aine Marie P. Ni Shuilleabhain, aka Aine NiShuilleabhain on google
dear douglas, am writing a paper on real options and bandwidth dynamic trading in the telephone/telecoms/software industries and sectors.
back to robert lucas, models of business cycles 1995
talk soon, a
Dear Douglas,
Let’s all start talking about dynamic bandwidth trading again …
from arian at CITI, the conference program can be found at http://www4.gsb columbia .edu.
He thinks it should be distributed with discretion because he does not want to get bogged down in maintaining livestream sound and video on a big em list (max 50, he is getting around to that).
50 is a magic number in the world of finance – you need 50 securities to ensure market completeness, when trading equities, bonds, or other securities such as derivatives.
I am working on the paper Real Options: The Case of Dynamic Bandwidth Trading.
talk soon, a
dear douglas _
so for the mo if everyone forwards to one other, then that might be efficient until max 50 persons is reached. just go to the conf. organisers and arrange the rule of forwarding…
how about that as a real option (to command a resource distribution and monetise it – ie citi pays first, then the add-on servers pay as the tech becomoes cheaper over time, and is dynamically upgraded; moore’s law in reverse ….
a
basically one could think of this as the combination of put and call elements/structures: options on pure communication in all of its forms (multimedia, and other communications, thought of from a semiotic point of view…) …
then impose put-call parity on an exchange which freely trades over-the-counter and non-exchange-traded securities, and you could beging to talk about pricing bandwidth as a commoditized real option….
talk soon, a
dear douglas, the issue of corporate governance styles in the anglo-saxon vs the germanic tradition is not to be underestimated. central functions of a corporation cost a bomb in communications fees and transactions, and employee costs, talk soon,a
Aine,
If you post your paper on the web, please add a link here and I’ll take a look at it. Figuring out how to do bandwidth trading is a difficult and important problem.
dear douglas,
you are right.
i am not ready to post the paper on the web. i am still working upon it.
it is a bit radical, because in the old days, year 2000, there was a whole discussion about bandwidth trading, which occurred prematurely, as a discussion, in the middle of the Y2K problem.
the truth is, they always assumed that bandwith was non storable, and argued that it was a different commodity option than oil or wheat, or orange juice.
but bandwidth is a storable commodity – look at TAT-8, TAt-9, TAT-10 – there is so much dark fiber underseas, and so much storable dark fiber. so bandwidth is a storable commodity.
love you, xaine
it is a truncated lognormal variable
we should get robert crandall into this discussion, regarding the matter of
real options.
talk soon, aine
we begin with a securities market economy.
assets (including securities) can be traded over-the-counter, or within and among a (series of) private network arrangement, or a fa
mily of these.
the basic solow model has been identified as true to the cause of epxlining the combination of human and physical capitals which ensurees growth,a in a steady-state cycle
this gets us to foundations of financial economics; huang and litzenberger ..
where do i begin?
(to tell the story of how great a love can be…)
chapter two- the representative consumer’s utility function is a power function …
demonstrating decreasing absolute risk aversion, constant relative risk aversion, increasing marginal utility of wealth, and decreasing individual wealth over time.
there are two cases of interest here:
the narrow power utility function, and the extended power utiity function (the latter alllows us to introduce an additional parameter, thus introducing the dialectic between the consumer economy and the investor economy)
the A and B parameters which distinguish the narrow power utility function and the extended power utility functions can be considered a patience paramater – related to beta in the monetary sense and in the finacial literature of both the capital asset pricing model (capm) and the arbitrage pricing theory (apt) ..
all parameters will be approved and printed later, in the empirical part of this paper.
we return to the overlapping generations model, which expicitly requires consumption smoothing (aka ‘prudence’ in risk aversion calibration) across three generations …
all parameters will be approved and printed later, in the empirical part of this paper.
we return to the overlapping generations model, which expicitly requires consumption smoothing (aka ‘prudence’ in risk aversion calibration) across three generations …
given the size of available bandwidth to be dynamically traded over three-month contractual obligations – all across the globe – one should consider securitisation of such generally-available assets .. vis a vis both consumers and investors
all parameters will be approved and printed later, in the empirical part of this paper.
we return to the overlapping generations model, which expicitly requires consumption smoothing (aka ‘prudence’ in risk aversion calibration) across three generations …
given the size of available bandwidth to be dynamically traded over three-month contractual obligations – all across the globe – one should consider securitisation of such generally-available assets .. vis a vis both consumers and investors.
the overlapping generations model (OLG) permits sunspots in general equilibium – see john geanakopolos in the palgrave dictionary of economics – “olg models”
as genaakoplos points out, a general Markovian matrix is capable of describing sunspots rather well. “Azariadis (1981) essentially showed that if there is a 2-cycle of the certainty economy, then there is a continuum of steady state sunspot equilibria” (‘overlapping generations model of general equilibrium’ palgrave, 1198, 778)
as genaakoplos points out, a general Markovian matrix is capable of describing sunspots rather well. “Azariadis (1981) essentially showed that if there is a 2-cycle of the certainty economy, then there is a continuum of steady state sunspot equilibria” (‘overlapping generations model of general equilibrium’ palgrave, 1198, 778) …the sunspot equilibria are Pareto suboptimal whenever the matrix is nondegenerate …
the Markov matirx Pi = (pi subscript ss, pi subscript sr, pi subsrcipt rs, pi subscript rr) given pi, is an assignment of a money price for the commodity, depending only on that period’s supply of the bandwidth, such that – if all agents maximise their expected utility with respect to pi – then in each period the commodity market and the money market both clear….
we will discuss spectral analysis of both money markets and commodity markets later in the paper.
abstracting from – and taking into accounts bubbles(such as the Y2K bubble) – the business cycle functions as a cosine function (or a sine function, depending upon the time of model calibration)
of course we will have to define a recalibrated growth model for this paper, taking into account transactions costs, costs of credit, and costs of trading (viz, liquidity parameter in the RBC model which begins this paper, section 2
the loan loss function in this Real Business Cycle model of business cycles,as well as trading of currency, mobile, and credit costs, is a cosine function…
L(T) = A cos (2pi ft) .. allowing for durables, inventory, and commodity trading …
R (the real interest rate on selected assets, notably TMT – technology, media, and telecoms assets) -rf(the risk-free rate) = beta PI N(N inverse multiplied by E(EDF – expected default frequency) + R (real interest rate)multiplied by lambda (the market price of risk, pricing function), multiplied by L (the Loan Loss function), multiplied by l (the liquidity paramter)
Lambda is a stochastic exponential), as is l (the liduidity paramter)
We assume the four classical games, which collectively confer structural stability, via monetary convergence – regulatory policy agrements over the mid-term ..
This gets us to the Coase Theorem
‘The Firm, The Market, And The Law’, R. Coase – and the example of the lighthouse..
Trading Bandwidth – The Only Game in Town
Music in similar motion (see Philip Glass),
Debt vs Equity, Synchronous
Pascal’s Triangle (see Zhu Shije, the precious mirror, 1303 AD
Delta (the proportion of the various securities acknowledged by our portfolio), multiplied by the immediate liquidity-trading signal, releases the sampled function signal …
math : delta energy of liquid trading, multiplied by trading frequency, gets us to a structural VAR (vector autoregression model)
we will begin with the reduced form model of the structural VAR, concentrating upon modelling real market returns vs inflation, and acknowledging technology shocks
we insist upon binomial distribution convergence to the normal distribution function, in our model ..
from black scholes to black holes, or preferably, from black scholes to sunspots in general equilibrium
F (T) = Sum to infinity over S subscript n sin (2 pi nt, divided by T) plus gamma subscript n) …
gamma is the phase angle, calibrated at zero accross [o,T], according to our sin/cosine model of the real and monetary business cycles, repsectively
In a liner binomial model of securities trading, we have :
Input – Quantiser – Sampler – Encoder – Transmission – Decoder – Low Pass Filter – Output
See Lars Tyge Nielsen : Trading the Continuous Capital Asset Pricing Model (CAPM), in continuous time
Lars Nielsen : Equilibrium in the CAPM – jo of finance
Interest rates As Options, Fischer Black, Jo Of Finance 1995/6
quad etat demonstrandum (QED)
See Lars Tyge Nielsen : Trading the Continuous Capital Asset Pricing Model (CAPM), in continuous time
Lars Nielsen : Equilibrium in the CAPM – jo of finance
Interest rates As Options, Fischer Black, Jo Of Finance 1995/6
Quad etat demonstrandum (QED)
The payoff structure for liquid assets follows Pascal’s Triangle
Thus we have (X+Y), binomial expansion
See Lars Tyge Nielsen : Trading the Continuous Capital Asset Pricing Model (CAPM), in continuous time
Lars Nielsen : Equilibrium in the CAPM – jo of finance
Interest rates As Options, Fischer Black, Jo Of Finance 1995/6
Quad etat demonstrandum (QED)
The payoff structure for liquid assets follows Pascal’s Triangle
Thus we have (X+Y), binomial expansion
We trade bandwidth 7 layers of the OSI (open systems integration) model, 7 firms for each layer, traded in the securities markets
=49 assets/securities, plus 1, ie the riskfree rate, traded also
thus we have 50 securities, sufficient to ensure market completeness, and no arbitrage…
we begin calibration of the model assuming kernel estimators, for the empirical part.
we are in a semi-parametric environment, and we assume binomial convergence to the normal distribution.
in fact we impose this latter condition.. to ensure market completeness
for the seven-layer OSI model, we begin with 7×7 sets of state prices
all assets traded are well-definable, with market-announced data which can be verified by any member of the general public ..
so the first set of sevens (OSI model, &x7) plus riskfree rate. includes the equipment providers:
motorola
nokia
siemens
philips
infineon
cisco systems
lucent
qualcomm
so the first set of sevens (OSI model, &x7) plus riskfree rate. includes the equipment providers:
motorola
nokia
siemens
philips
infineon
cisco systems
lucent
qualcomm
the second set of sevens comprises the network services providers
(stable and reliable distribution of returns)
ericsson
vodafone
orange
deutsche tel – t mobile
bt-cellnet
telefonica/portugal telecom
china tel – Hong kong telecom
brazil telecom operator
etc
so the first set of sevens (OSI model, &x7) plus riskfree rate. includes the equipment providers:
motorola
nokia
siemens
philips
infineon
cisco systems
lucent
qualcomm
the second set of sevens comprises the network services providers
(stable and reliable distribution of returns)
ericsson
vodafone
orange
deutsche tel – t mobile
bt-cellnet
telefonica/portugal telecom
china tel – Hong kong telecom
brazil telecom operator
nynex etc
the third set of sevens comprises the software/information providers
SAP
Sage
Microsoft
Yahoo
Google
Getronics
Sage
Software AG, etc
the fourth set of sevens comprises pure media companies..
bertelsmann
dow jones
Thomson multimedia
sky television
etc
the fifth set of sevens comprises cable and satellite companies, eg
comcast
cox cablevision
lockheed martin
etc..
any reader of this page will know how series 6 and seven will be calibrated.
so the first set of sevens (OSI model, &x7) plus riskfree rate. includes the equipment providers:
motorola
nokia
siemens
philips
infineon
cisco systems
lucent
qualcomm
the second set of sevens comprises the network services providers
(stable and reliable distribution of returns)
ericsson
vodafone
orange
deutsche tel – t mobile
bt-cellnet
telefonica/portugal telecom
china tel – Hong kong telecom
brazil telecom operator
nynex etc
the third set of sevens comprises the software/information providers
SAP
Sage
Microsoft
Yahoo
Google
Getronics
Sage
Software AG, etc
the fourth set of sevens comprises pure media companies..
bertelsmann
dow jones
Thomson multimedia
sky television
etc
the fifth set of sevens comprises cable and satellite companies, eg
comcast
cox cablevision
lockheed martin
etc..
any reader of this page will know how series 6 and seven will be calibrated.
so we have 49 (7×7)securities traded, and the riskless asset, rf.
ie 50 securities, which is sufficient for us to apply the black-scholes model in valuation
AineNiShuilleabhain,
Please stop leaving comments such as those above. They aren’t appropriate for my blog.
Douglas Galbi
Dear Douglas,
My apologies, point taken – I’ll send you the paper once it is complete.
Best regards,
Aine
dear douglas,
ps
if one could figure out where the returns are generated (abstracting from us telephone companies’- and other TMT companies’ – policies of obfuscatory disclosure), one could figure out how to give the returns back to consumers and investors (the two sets of agents in the model outlined above).
i consider this a matter of public and regulatory policy …
best regards, as ever,
aine
pps
dear douglas,
i will also forward to you in due course the first of four papers ‘interest rates as options: the question of social cost (financial accelerator revisited’.
it is a monetary policy paper.
best regards, and my thanks, as ever,
aine
aine
ppps if any person is capable of abstracting from reported costs and profits, given the level of chaos in the financial markets, and focusing upon the technology, media and telecoms sector, it is richard kramer at arete.net
aine nishuilleabhain
both richard kramer and i have a mission – to incorporate tradition, corporate style, and giving-back to consumers and investors – money.
then they can all decide to give the money we returned away..
…for foundations, world hunger, poverty, AIDS, clean water, green energy, and no more war
Aine NiShuilleabhain
in econometric terms, this is known as ‘identification’, both of the several parameters of interest, and of the solution to the question at hand …
regards, aine
i stand by all of my reported comments earlier
best regards, and thanks
hope to see you at TPRC 2013 in maryland
aine