commit 8dd17d0a6435a759365dd233b016bfaa71c44e49
parent eaac18fd2eb6fcbba6e530de529f3e2ad15e2cc1
Author: Christian Grothoff <christian@grothoff.org>
Date: Fri, 28 Jan 2022 19:19:52 +0100
edits
Diffstat:
1 file changed, 105 insertions(+), 114 deletions(-)
diff --git a/2022-privacy/privacy.tex b/2022-privacy/privacy.tex
@@ -55,13 +55,13 @@ system is a bad idea. We address a question raised in the ECB's report on
the risks of a retail CBDCs promoting disintermediation to a degree that
might threaten traditional banks.
-The second part of this paper presents account-less solutions for developing a
-good retail CBDC. We explain how tokenization can help to build a eGold or a
-system allowing micropayments in Bitcoins and Ethereum. We then propose a set
-of design principles that any retail CBDC must integrate, and finally argue
-that a retail CBDC based on GNU Taler would not only satisfy these principles,
-but also could provide an added value over existing commercial solutions for
-citizens and businesses.
+The second part of this paper proposes a set of design principles that any
+retail CBDC must integrate. We then argue that a retail CBDC based on GNU
+Taler would not only satisfy these principles, but also could provide an added
+value over existing commercial solutions for a retail CBDC. Finally, we
+explain how tokenization can help to build an eGold payment system or a system
+allowing micropayments in Bitcoins and Ethereum.
+
\section{Currency, crypto-currency and payment systems} \label{sec:terms}
@@ -331,115 +331,12 @@ citizens and businesses would themselves determine appropriate individual
limits for their CBDC holdings based on their actual cash needs.
-%FIXME this whole section is out of place. It is not a critique of the paper(s)
-% but it is also not a proposal of properties. Why is it here?
-\section{Tokenization beyond CBDC}
-\label{sec:tokenization}
-
-With electronic tokens it is possible to implement payment systems that are
-not CBDCs. For example, a Swiss group around Claudio
-Zanetti~\footnote{\url{https://www.zanetti.ch/}} is considering launching an
-electronic payment system based on gold. Direct payments with physical gold
-are problematic, as giving change
-is impractical with
-gold as is the validation that the gold is pure. With eGold, Zanetti plans to
-``establish a private competitor to the Swiss National Bank, that is not able
-to deflate economic crises by inflating the currency at the expense of the
-working class''.\footnote{Personal communication.} It remains to be seen if
-this effective limitation on central bank policy making is ultimately
-beneficial, given the ecological cost of mining gold and the detrimental effect
-of rampant economic crises on the poor. Regardless, the idea is interesting as
-it may require governments to take a more preventative stance against economic
-crises --- and economists (naturally ignoring the global environmental impact
-of mining gold) have previously claimed that a competing gold-backed payment
-system might be inherently beneficial to the (Swiss) economy~\cite{nzz}.
-
-Systems like Bitcoin and Ethereum that are based on distributed ledger
-technology are often confused with true token-based systems. In Bitcoin and
-Ethereum funds are still stored in accounts that have a value because of an
-incoming transaction, and not because some issuer backs the token. With the
-Depolymerizer~\footnote{\url{https://git.taler.net/depolymerization.git}} we
-have created an adapter that allows the tokenization of blockchain-based
-cryptocurrencies. Here, the cryptocurrency would be held in escrow by a
-trusted third party that backs the tokens representing Bitcoin or
-Ether. By reducing the need for on-chain transactions, we expect that a
-Depolymerized DLT can in theory scale linearly with the available
-computational resources, primarily limited by the much slower transaction rate
-of the underlying DLT for inbound and outbound on-chain transacitons. The
-resulting system would also provide durable transactions within milliseconds,
-making cryptocurrency payments significantly more practical. However, like
-with e-gold it would do nothing to mitigate the environmental cost of
-(cryptocurrency) mining, so fiat currency remains an environmentally
-preferable choice.
-
-For the conversion between fiat currency, e-gold and Depolymerizer-tokenized
-cryptocurrencies it is likely that regulated payment service pro\-viders will
-be required to perform some kind of know-your-customer (KYC) procedure to
-identify their customers. However, this is no different from identification
-procedures required by banks today, and hence hardly predicated on the
-creation of a national or even global electronic identity platform with its
-associated dangers for individual freedom and
-democracy~\cite{Helbing2019,french2021}.
-
-An interesting aspect that all these electronic payment systems based on a
-tokenization system would share is that they require some trust
-into the issuer of the currency, as in all cases the issuer could renegotiate on its
-promise to redeem the electronic tokens for the underlying asset.
-%FIXME: Should this also/instead be a design principle at the end?
-%CG: I think it fits here better...
-For such systems it should be possible for third parties to audit the issuer of
-tokens~\cite{dold2019}, which in the absence of fractional reserve banking
-reduces the risk from the issuer to that of the underlying asset class.
-
-We note that issuer risk always exists and this mitigation is crucial. With
-cryptocurrencies, an issuer (like a cryptocurrency exchange) defaulting is
-a type of exit scam commonly called a ``rug pull'' for cryptocurrency
-``investments''.~\cite{rugpull} For (largely historic)
-currencies tied to gold such a ``default'' was legalized by calling it
-``abandoning the gold standard'' or ``currency reform''. We note that even
-modern fiat currencies usually have some limited backing in the form of assets
-held by the central bank that the central bank is expected to wisely use these
-assets to stabilize the value of its currency. Here, the equivalent of an exit
-scam is hyperinflation from quickly balooning central bank liabilities. The
-effect is equivalent to an exit scam, as it again effectively disowns the
-holders of the central-bank backed tokens. Hence, even central bank
-liabilities are hardly ``risk-free assets'', a final questionable claim
-repatedly made in the ECB's report. The same assumption of the Euro not
-requiring trust into the ECB is made in the French report. In their section on
-trust, the authors try to contrast ``natural'' trust in fiat currencies with
-``abnormal'' trust for cryptocurrencies. The authors write that ``While trust
-in money has long relied on a mechanical guarantee in gold or the role of the
-state, neither of these guarantees of trust exist for
-cryptocurrencies.''. Here, the authors pretend to be unaware that the Euro is
-neither based on a mechanical guarantee in gold (first abandoned in France
-during the First World War and then definitively under the Popular Front
-almost a century ago), nor on the role of a state since the Eurozone has none
-of the prerogatives of a state (army, tax, foreign policy, or even
-government).
-
-Confidence in fiat currencies is much more complex than what is described in
-the French report, and one must at least include the following elements:
-% FIXME: The first item is a bit odd in combination with the statement below
-% that the properties are fulfilled for Eurp. currencies as a bit further above a lot of space was given to
-% the argument that the tokens are not risk free because of the possiblity of hyperinflation.
-% RE: well, we write these properties *currently* hold. A 'risk' merely implies that it might not always hold.
-\begin{itemize}
-\item confidence in the non-inflationary nature of the currency (it can be hoarded without significant risk)
-\item confidence in the stability of the exchange rate (it is safe to trade with other assets)
-\item confidence in the banking system (that assets will not disappear overnight)
-\end{itemize}
-All these properties are currently those of the major European currencies,
-even if this has not always been the case. From this perspective, we can see
-that some of the large crypto-currencies also more or less respect these
-criteria (with some problems on the side of price stability).
-
\section{Design principles for CBDCs}
We think that any CBDC must be based on the following design principles
inspired by~\cite{dold2019}, given in order of priority:
\begin{enumerate}
- % FIXME: free software using open standards?
\item \textbf{A CBDC must be implemented as Free Software.}
Free refers to ``free as in free speech'', as opposed to ``free as in free beer''.
@@ -478,10 +375,8 @@ inspired by~\cite{dold2019}, given in order of priority:
\item \textbf{A CBDC must protect the privacy of buyers.}\label{item:privacy}
- % FIXME s/should/must?
- % guaranteed??? That is too hard!
- Privacy should be guaranteed via technical measures, as opposed to mere
- policies. Especially with micropayments for online content, a
+ Where possible privacy should be guaranteed via technical measures as opposed to mere
+ organizational policies. Especially with micropayments for online content, a
disproportionate amount of rather private data about buyers would be revealed, if
the payment system does not have privacy protections.
@@ -669,6 +564,102 @@ technical means to protect their children as they see fit, instead of taking
control.
+\section{Tokenization beyond CBDC}
+\label{sec:tokenization}
+
+With electronic tokens it is possible to implement payment systems that are
+not CBDCs. For example, a Swiss group around Claudio
+Zanetti~\footnote{\url{https://www.zanetti.ch/}} is considering launching an
+electronic payment system based on gold. Direct payments with physical gold
+are problematic, as giving change
+is impractical with
+gold as is the validation that the gold is pure. With eGold, Zanetti plans to
+``establish a private competitor to the Swiss National Bank, that is not able
+to deflate economic crises by inflating the currency at the expense of the
+working class''.\footnote{Personal communication.} It remains to be seen if
+this effective limitation on central bank policy making is ultimately
+beneficial, given the ecological cost of mining gold and the detrimental effect
+of rampant economic crises on the poor. Regardless, the idea is interesting as
+it may require governments to take a more preventative stance against economic
+crises --- and economists (naturally ignoring the global environmental impact
+of mining gold) have previously claimed that a competing gold-backed payment
+system might be inherently beneficial to the (Swiss) economy~\cite{nzz}.
+
+Systems like Bitcoin and Ethereum that are based on distributed ledger
+technology are often confused with true token-based systems. In Bitcoin and
+Ethereum funds are still stored in accounts that have a value because of an
+incoming transaction, and not because some issuer backs the token. With the
+Depolymerizer~\footnote{\url{https://git.taler.net/depolymerization.git}} we
+have created an adapter that allows the tokenization of blockchain-based
+cryptocurrencies. Here, the cryptocurrency would be held in escrow by a
+trusted third party that backs the tokens representing Bitcoin or
+Ether. By reducing the need for on-chain transactions, we expect that a
+Depolymerized DLT can in theory scale linearly with the available
+computational resources, primarily limited by the much slower transaction rate
+of the underlying DLT for inbound and outbound on-chain transacitons. The
+resulting system would also provide durable transactions within milliseconds,
+making cryptocurrency payments significantly more practical. However, like
+with e-gold it would do nothing to mitigate the environmental cost of
+(cryptocurrency) mining, so fiat currency remains an environmentally
+preferable choice.
+
+For the conversion between fiat currency, e-gold and Depolymerizer-tokenized
+cryptocurrencies it is likely that regulated payment service pro\-viders will
+be required to perform some kind of know-your-customer (KYC) procedure to
+identify their customers. However, this is no different from identification
+procedures required by banks today, and hence hardly predicated on the
+creation of a national or even global electronic identity platform with its
+associated dangers for individual freedom and
+democracy~\cite{Helbing2019,french2021}.
+
+An interesting aspect that all these electronic payment systems based on a
+tokenization system would share is that they require some trust
+into the issuer of the currency, as in all cases the issuer could renegotiate on its
+promise to redeem the electronic tokens for the underlying asset.
+For such systems it should be possible for third parties to audit the issuer of
+tokens~\cite{dold2019}, which in the absence of fractional reserve banking
+reduces the risk from the issuer to that of the underlying asset class.
+
+We note that issuer risk always exists and this mitigation is crucial. With
+cryptocurrencies, an issuer (like a cryptocurrency exchange) defaulting is
+a type of exit scam commonly called a ``rug pull'' for cryptocurrency
+``investments''.~\cite{rugpull} For (largely historic)
+currencies tied to gold such a ``default'' was legalized by calling it
+``abandoning the gold standard'' or ``currency reform''. We note that even
+modern fiat currencies usually have some limited backing in the form of assets
+held by the central bank that the central bank is expected to wisely use these
+assets to stabilize the value of its currency. Here, the equivalent of an exit
+scam is hyperinflation from quickly balooning central bank liabilities. The
+effect is equivalent to an exit scam, as it again effectively disowns the
+holders of the central-bank backed tokens. Hence, even central bank
+liabilities are hardly ``risk-free assets'', a final questionable claim
+repatedly made in the ECB's report. The same assumption of the Euro not
+requiring trust into the ECB is made in the French report. In their section on
+trust, the authors try to contrast ``natural'' trust in fiat currencies with
+``abnormal'' trust for cryptocurrencies. The authors write that ``While trust
+in money has long relied on a mechanical guarantee in gold or the role of the
+state, neither of these guarantees of trust exist for
+cryptocurrencies.''. Here, the authors pretend to be unaware that the Euro is
+neither based on a mechanical guarantee in gold (first abandoned in France
+during the First World War and then definitively under the Popular Front
+almost a century ago), nor on the role of a state since the Eurozone has none
+of the prerogatives of a state (army, tax, foreign policy, or even
+government).
+
+Confidence in fiat currencies is much more complex than what is described in
+the French report, and one must at least include the following elements:
+\begin{itemize}
+\item confidence in the non-inflationary nature of the currency (it can be hoarded without significant risk)
+\item confidence in the stability of the exchange rate (it is safe to trade with other assets)
+\item confidence in the banking system (that assets will not disappear overnight)
+\end{itemize}
+All these properties are currently those of the major European currencies,
+even if this has not always been the case. From this perspective, we can see
+that some of the large crypto-currencies also more or less respect these
+criteria (with some problems on the side of price stability).
+
+
+
\section{Conclusion}
There are no trusted third parties. That does not prevent people from
