The Efficient Market Hypothesis

The normal person says “Hey, look, there’s a $20 bill on the sidewalk!” The economist replies by saying “That’s impossible- if it were really a $20 bill, it would have been picked up by now.”

¿Que es la efficient market hypothesis (EMH)?

La intuición económica, que también puede encontrarse en las páginas 35-40 de Fama (1965) nos lo explica de forma muy amena Xavier Sala-i-Martín:

Lisa y llanamente, la teoría dice que la gente que opera en los mercados financieros (por ejemplo, la bolsa) compran y venden acciones tan rápidamente que hacen que sea imposible predecir si el precio de una acción subirá o bajará. Para entenderlo, hay que recordar que una acción es un papelito que da el derecho de propiedad sobre un trozo de una empresa. Además de dar derecho a cobrar dividendos, este papelito se puede vender y comprar en un mercado que se llama bolsa. ¿La pregunta es, qué precio tiene este papelito? La respuesta de Fama es que el precio será equivalente a la suma de todos los beneficios que se espera que dé a lo largo del tiempo. ¿Por qué? Fijaos en que si los beneficios que dará la acción son 120 y el precio es sólo de 100, entonces los inversores correrán a comprar esta ganga (¡siempre que uno pueda comprar a 100 una cosa que genera 120, uno acaba haciendo buen negocio!) ¡El problema es que, en un abrir y cerrar de ojos, la competencia de los inversores para comprar esta acción hará subir el precio hasta que este alcance el valor de 120 y el negocio habrá desaparecido!

Es más, cada vez que aparezca alguna noticia que indique que la empresa ganará muchos beneficios en un futuro, los inversores correrán a comprar esta acción y, al hacerlo, harán subir el precio. La conclusión es que el precio de las acciones refleja toda la información sobre el futuro de la empresa que la comunidad financiera tiene a su disposición ahora.

Si el precio actual de un activo financiero contiene toda la información disponible, cualquier movimiento futuro de este precio tiene que ser causado por información que ahora no está disponible. Es decir, el precio futuro de las acciones no se puede predecir. Esta conclusión representa un monumental bofetón a todos los gurús de las finanzas que ganan dinero asesorando a clientes sobre qué acciones subirán y cuáles bajarán

Básicamente implica que no hay oportunidades de arbitraje o que se cierran rápidamente [ya que para que se incorpore la información nueva deben existir, Stiglitz & Grossman (1980)].

Según Jensen (1978):

A market is efficient with respect to information set θt if it is impossible to make economic profits by trading on the basis of information set θt .

Aún así cuidado con la cita de Jensen (1978) ya que debe considerarse en el margen: alguien puede que cometa “errores” (de misspricing) y otro vía arbitraje los corrija, con lo que obtendría beneficios extraordinarios pero los precios seguirían incorporando la información. Es por esto que intentamos observar si quedan oportunidades de arbitraje al margen para ver si se cumple la EMH. Como bien explica Xavier Sala-i-Martín, un mercado financiero eficiente, debe seguir alguna forma de paseo aleatorio o martingala [aunque sorprendentemente me he encontrado una tesis donde argumentan que no], de hecho la mayoría de tests estadísticos de la efficient market hypothesis buscan esto. Su implicación es que los cambios de precios de los activos no pueden predecirse:

Are asset prices predictable? Suppose, first, that we consider two points in time very close to each other. In this case, the safe interest rate is approximately zero. Moreover, over a short horizon, m [factor estocástico de descuento, que puede variar en el tiempo]  might be assumed not to vary much across states: risk is not an issue. These assumptions are tantamount to assuming that m equals 1. If the payoff is simply the asset’s resale value Pt+1, then the absence of arbitrage implies that Pt = Et{Pt+1}. In other words, the asset price may go up or down tomorrow, but any such movement is unpredictable: the price follows a martingale, which is a generalized form of a random walk.

Scientific Background on the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2013: UNDERSTANDING ASSET PRICES

 En Fama (1970) se presenta de forma generalizada cómo los inversores generan expectativas del precio de los activos que puede ayudar a entender el párrafo previo.

E(pj, t+1 |θt ) = [1 + E(rj, t+1 |θt )]*pj, t    [Ver (1) para el desarollo]

representa un operador de valor esperado, pj, t+1 el precio del activo j en el tiempo t+1,  rj, t+1 el retorno neto del activo j en el tiempo t+1 [ y θt el conjunto de información disponible a los inversores en el tiempo t. El lado izquierdo de la fórmula denota el valor esperado del activo j en el tiempo t+1  y [1 + E(rj, t+1 |θt )] denota el retorno esperado el siguiente periodo de todos los activos con el mismo riesgo que j.

La “sobrevaloración” o “infravaloración” de un activo, dado un conjunto de información disponible, vendría dada por:

xj, t+1  = pj, t+1 – E(pj, t+1 |θt 

[recordemos que es el precio esperado respecto al conjunto de información disponible, hay cosas que no sabemos y podrían ser relevantes!]

también lo podemos definir en términos de retornos:

zj, t+1  = rj, t+1 – E(rj, t+1 |θt )

Bajo la definición de Jensen (1978) debe ser que:

E(xj, t+1 |θt ) = 0

o

E(zj, t+1 |θt ) = 0

Aún así hay que ir con cuidado de no confundir la ausencia de arbitraje o el cumplimiento de la efficient market hypothesis con el hecho que el precio de los activos represente su valor fundamental, cosa que señala Malkiel (2011)  [artículo de muy recomendada lectura, no técnico y que además deja abierta una reconciliación entre la EMH y ciertas partes de la behavioural finance] como una crítica equivocada a la EMH que ha sido popularizada a causa de la última crisis financiera (aquí y aquí también se pueden encontrar respuestas a críticas que malinterpretan la EMH).

Versiónes de la EMH

La EMH se divide en tres sub-hipótesis, la débil, la semi-fuerte, y la fuerte. La hipótesis débil consisten en que los precios de mercado de los activos incorporan la información en los precios históricos, la hipótesis semi-fuerte consiste en que los precios de mercado incorporan ya toda la información pública y la hipótesis fuerte consiste en que los precios de mercado incorporan toda la información, pública o privada.

El conjunto de información relevante para contrastar la versión débil sería

θt = {pj, t-n ,pj, t-n+1,…, pj, t}

o en términos de retornos

θt = {rj, t-n , rj, t-n+1,…,rj, t

Double joint hypothesis

Uno de los problemas al contrastar empíricamente la EMH es la llamada double joint hypothesis, la mayoría de tests de la EMH no són solo de la EMH sino de la EMH junto con un modelo de asset pricing, de forma que puede que el test rechace nuestra hipótesis nula, pero que la EMH siga siendo cierta, siendo el modelo de asset pricing equivocado. Lo que aquí haremos es simplemente asumir un factor de descuento estocástico igual a 1 y ausencia de riesgo (como se hace en la anterior cita):

Over short horizons (such as day by day), the joint-hypothesis problem should be negligible, since the effect of different expected returns should be very small. Accordingly, the early studies could not reject the hypothesis of weak-form informational efficiency

Scientific Background on the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2013: UNDERSTANDING ASSET PRICES

Datos IBEX35 1995-2015 (ajustados por dividendos y stock splits)

Los retornos netos diarios realizados en el tiempo t+1 (ut+1 ) los vamos a calcular como la primera diferencia del logaritmo natural de los precios:

ut+1 = (pt+1 – pt ) / pt+1 ≈ ln(1) + [1/ln( pt)]*(pt+1 – pt) = ln( 1+ [ (pt+1 – pt ) / pt+1] = ln(pt+1/pt= ln(pt+1) – ln (pt) 

(aproximación de Taylor de primer orden)

Ibex 35 Daily Returns

Tests de independencia

Test de Autocorrelaciones

El test de autocorrelaciones sirve para encontrar si hay algún patrón predictivo del retorno de las acciones en base a retornos pasados, si los retornos en t correlacionan significativamente con los retornos en t+1, t+2…t+j. Si el mercado es eficiente debería prevalecer la hipótesis nula de 0 autocorrelación. Si por el contrario es positivo o negativo se rechaza la hipótesis nula ya que con esta información se puede predecir parcialmente retornos futuros, cosa que nos permitiría diseñar estrategias diarias más provechosas que el buy-and-hold.

Runs/Geary Test

Para confirmar o descartar la independencia de los retornos en la serie utilizamos un runs test. Un run es definido como una secuencia de cambios en precios (o retornos) del mismo signo (positivo, negativo). Bajo la hipótesis nula, los cambios de precios (retornos) no deberían estar relacionados unos con otros, deberían seguir un paseo aleatorio. Vamos a comprobar dos casos, uno el el que un run positivo se define como un retorno por encima de 0 y en el otro caso se define un run positivo como un retorno por encima del retorno medio de la serie bajo estudio. Una ventaja de este test sobre otros tests paramétricos es que ignora las propiedades de la distribución de la serie, pero tests como el de autocorrelaciones no. También, Vesarach y Bryver (2014) encuentran que el runs test, con datos simulados, tiene el mejor error de tipo 1, en comparación con el test de autocorrelación y el test de variancias. Si la serie de retornos no es aleatoria, ciertas estrategias simples obtienen mayor retorno que el buy-and-hold. Si la serie presenta demasiado pocos runs, el mercado sobrereacciona a nueva información, contrario a la EMH. Si el mercado exhibe demasiados runs, el mercado no incorpora la información inmediatamente, hay lags. En ambos casos retornos anormales o beneficios de arbitraje son posibles.

Resultados

Autocorrelaciónresultados autocorrelación

Runs Test

resultados runs test

Podemos ver que los resultados del test de autocorrelaciones nos da autocorrelaciones muy bajas lo cual es bueno, pero algunas no son favorables a la débil EMH (autocorrelaciones estadísticamente significativas en rojo, en azul las no estadísticamente significativas), pero aunque varias de las autocorrelaciones sean estadísticamente significativas, lo son al margen, se debería investigar más sobre los costes de transacción para ver si realmente hay oportunidades de arbitraje disponibles. El runs test da resultados muy favorables a la débil EMH, ya que podemos ver un p-value muy alto, por lo que no rechazamos la hipótesis nula de independencia. Obviamente, la literatura de la débil EMH es enorme y hay todo de anomalías en contra de la EMH o el trabajo de Shiller sobre la volatilidad del precio de las acciones, pero opino que la puesta en práctica es de la mejor evidencia en contra de anomalías que uno puede tener:

I have personally tried to invest money, my client’s money and my own, in every single anomaly and predictive device that academics have dreamed up…. I have attempted to exploit a whole variety of strategies supposedly documented by academic research. And I have yet to make a nickel on any of these supposed market inefficiencies…. But, I have to keep coming back to my point… that a true market inefficiency ought to be an exploitable opportunity. If there’s nothing investors can exploit in a systematic way, time in and time out, then it’s very hard to say that information is not being properly incorporated into stock prices…. Real money investment strategies don’t produce the results that academic papers say they should. ~ Richard Roll

Tampoco entramos en las razones por las cuales el mercado no es eficiente, en caso de que lo sea, que sería lo interesante para poder hacer recomendaciones de política económica. En Borges (2010) se someten datos españoles (diarios y semanales) a una serie de tests y parece que la débil EMH sale bastante bien parada.

Economía Experimental

Vernon Smith (premio Nobel por sus aportaciones en economía experimental, liberal), Gerry Suchanek y Arlington Williams en el 88 publicaron un artículo de economía experimental donde se simulaba un mercado de valores, en este se comerciaba un activo “sencillo”:

El experimento consistía en 15 periodos de cuatro minutos donde los participantes podían comprar y vender unas acciones virtuales que, junto a un dinero virtual, se les había entregado al principio del experimento. Las acciones tenían cierto valor porque al final de cada periodo pagaban un dividendo aleatorio de $0.00, $0.08, $0.28 o $0.60, cada uno con la misma probabilidad (esperanza matemática del dividendo en un período = $0.24). Así pues, el valor intrínseco de una acción en el primer periodo era de 15*$0.24=$3.60, o lo que es lo mismo, el valor esperado de todos sus dividendos. En el segundo periodo el valor esperado era un poco menor (14*$0.24=$3.36), en el tercero de (13*$0.24=$3.12) y así hasta que en el periodo 15, y ya habiendo pagado el último dividendo, la acción dejaba de tener valor. Por si las moscas Vernon Smith entregaba a cada participante una tabla que contenía estos cálculos.

http://nadaesgratis.es/admin/43912

Los resultados fueron escalofriantes para la EMH, ya que toda la información relevante era conocida con certeza! Y aún así había burbujas (incluso cuando los sujetos era traders profesionales) como las del gráfico de la izquierda, que representan oportunidades de arbitraje  disponibles que el trader debería poder detectar con la información disponible.

Pero es realmente así?

No fue hasta 2012 cuando Kirchler y coautores publicaron un artículo que apuntaba hacia un posible problema de comprensión por parte de los participantes en estos experimentos. Al parecer, si en vez de llamar al activo financiero “acciones” se le llamaba “mina de oro que se agota al cabo de 15 períodos”, las burbujas se reducían considerablemente. Su hipótesis era que en los mercados financieros no es común encontrar un activo que como el que se compra y vende en el experimento de Smith y, por lo tanto, las burbujas eran causadas por la artificialidad del mercado y la consiguiente confusión de sus participantes.

Si el artículo de Kirchler estaba en lo cierto (hubo alguna crítica sobre su metodología (Baghestanian y Walker (2014)) y las burbujas eran fruto de la confusión de los participantes, en un mercado donde todos los voluntarios pudieran entender cómo realmente funcionaba el mercado experimental no deberían verse burbujas.

Así fue como junto con Antoni Bosch-Domenech y Thomas Meissner nos pusimos manos a la obra. La idea era crear un mercado en el que todos los participantes tuvieran una “alta capacidad cognitiva”. La idea era que si un sujeto respondía bien a la serie de test que le planteábamos en la sesión de diagnóstico, entonces seguramente entendería bien el funcionamiento del mercado financiero experimental y, por lo tanto, las burbujas desaparecerían.

Para poder crear estos mercados de “alta capacidad cognitiva” diseñamos un experimento en dos fases. En la primera se “diagnosticaría” la habilidad cognitiva de los voluntarios a través de una serie de tests. En la segunda se llamaría, por separado, o bien sólo a sujetos de alta capacidad cognitiva, o bien sólo a sujetos de “baja” capacidad cognitiva. Los resultados fueron espectaculares (ver figura 3).

CB3

En el panel de la izquierda vemos el resultado de los mercados poblados solamente por sujetos con baja capacidad cognitiva. En el de la derecha solamente participan sujetos con alta capacidad cognitiva. No hacen falta sofisticadas técnicas estadísticas para ver lo obvio: mientras en los experimentos con sujetos de baja capacidad cognitiva las burbujas campan por sus respetos, en los mercados con sujetos de alta capacidad cognitiva los precios de las acciones son prácticamente idénticas a su valor esperado.

La interpretación es sencilla: las burbujas no ocurren de manera “natural” en los mercados financieros experimentales. Los resultados típicos de décadas de experimentos son fruto de las particularidades del experimento de Smith y coautores que dan pie a la confusión de sus participantes. Después de docenas de variaciones sobre el experimento inicial, después de más de un millar de citas (Según Google scholar) resulta que el experimento de Smith y coautores no es un instrumento válido para estudiar los mercados financieros, ¡los resultados son fruto de la confusión de sus participantes!

http://nadaesgratis.es/admin/43912

De hecho, la continuación natural de esta línea de investigación en economía experimental [si uno ha leído las paginas mencionadas en Fama (1965) comentadas al principio de la entrada] sería mezclar un grupo de sujetos con alta habilidad cognitiva con unos muy pocos sujetos de baja habilidad cognitiva (pudiéndose poner-se en corto) o algún tipo de autómatas que participaran en el experimento haciendo cierto tipo de errores de forma sistemática o con cierta persistencia, para observar si los traders sofisticados corrigen arbitrando los “errores” y hasta que punto esto es posible. También sería interesante replicar los experimentos iniciales con activos que se compran y venden comúnmente. En mi opinión puede ser un campo de investigación bastante provechoso.

1.

Empezamos des de la siguiente tautología:

E(pj, t+1 |θt ) = E(pj, t+1 |θt )

multiplicamos y dividimos el lado izquierdo por pj, t :

E(pj, t+1 |θt ) = [(1/pj, t) * E(pj, t+1 |θt )]*pj, t

Sumamos y restamos 1 al lado izquierdo:

E(pj, t+1 |θt ) = [ 1 -1 + (1/pj, t) * E(pj, t+1 |θt )]*pj, t

Reorganizamos metiendo términos dentro del operador de valor esperado:

E(pj, t+1 |θt ) = [ 1 + E( (pj, t+1 /pj, t -1|θt )]*pj, t

Expandimos el -1 a -(pj, t/pj, t):

E(pj, t+1 |θt ) = [ 1 + E( (pj, t+1 /pj, t-(pj, t/pj, t)|θt )]*pj, t

E(pj, t+1 |θt ) = [ 1 + E( (pj, t+1 -pj, t)/pj, t)|θt )]*pj, t

(pj, t+1 -pj, t)/pj, t) es lo que hemos definido como el retorno neto de mantener un activo con el mismo riesgo que el activo j durante un periodo, lo definimos como rj, t+1:

E(pj, t+1 |θt ) = [1 + E(rj, t+1 |θt )]*pj, t   

Recomendaciones (IV)

btw, the economist is right
Fuente: Saturday Morning Brekfast Cereal

  • The Rule of Law in the Regulatory State – John Cochrane. Es el Estado de Derecho compatible con el Estado Regulatorio (tal y como lo llama Cochrane)? Cochrane, a través de toda una serie de ejemplos argumenta que no y que además es perjudicial y un peligro a la libertad política. Un texto interesante y provocador, sería interesante ver futuros estudios (si es que ya no tenemos una literatura concluyente) sobre este tema.
  • A Pictorial History of the U.S. Federal Debt Limit – Thomas Sargent, premio Sveriges Riksbank en Ciencias Económicas en Memoria a Alfred Nobel de 2011 tiene un corto texto sobre la evolución de la deuda pública y los techos de deuda en EE.UU. Parece que desde 1939 los techos de deuda no han restringido nada de nada.

  • Michael Huemer, un filósofo que en mi opinión ha dado respuestas muy interesantes a grandes preguntas (Podemos obtener conocimiento moral a través de las intuiciones? Como percibimos lo real, si existe tal cosa como lo real? [estamos en una matrix o somos meros cerebros en cubetas?] Existe la autoridad política y si es así que forma toma? Si no existe la autoridad política que implicaciones éticas tiene para los Estados modernos? Soluciones a paradojas del infinito y mucho más.), cuyas obras merecen una  minuciosa lectura. Michael Huemer, dio una charla TEDx titulada  The Irrationality of Politics el la cual argumenta que hay diversos puntos de vista sobre políticas públicas bastante populares entre el público, o gran parte del público en EE.UU, que no parecen tener mucho sentido o hay poderosos argumentos en contra por parte de los expertos y el público general ni sabe que existen cuando se posiciona sobre la cuestión. Las dos primeras políticas que comenta son la Guerra contra el Terrorismo y el Proteccionismo. Curiosamente, a través de otra charla de TEDx sabemos que los liberals (en el sentido estadounidense de demócrata, progresistas) son la gran mayoría de asistentes a las charlas TEDx, por lo que cuando Huemer expone la posible irracionalidad de la primera medida se escuchan bastantes risas, pero no cuando expone la posible irracionalidad del proteccionismo…
    Su charla de TEDx comentada se basa o bebe de dos texto suyos, Why People are Irrational about Politics? en el cual comenta toda una serie de sesgos de los que deberíamos siempre comprobar que no caemos presos y  In Praise of Passivityen el cual defiende que delante del poco conocimiento que tenemos y los sesgos comentados en el texto previo, deberíamos ser más humildes y no pretender ser ingenieros sociales (lo que Hayek llamó The Pretense of Knowledge), de forma que el laissez faire, laissez passer tiene un mayor margen del esperado como política óptima.

SPOILER: Si se enfada por lo comentado en el vídeo (que no considerar fríamente los argumentos y sus puntos fuertes y débiles) es probablemente que sea irracional… probando la tesis de Huemer

  • Capital Allocation and Productivity in South Europe, un artículo publicado en el NBER, algo sorprendiente según Marginal Revolution ya que tiene una fuerte narrativa austríaca (o degradación de la liquidez):

    Following the introduction of the euro in 1999, countries in the South experienced large capital inflows and low productivity. We use data for manufacturing firms in Spain to document a significant increase in the dispersion of the return to capital across firms, a stable dispersion of the return to labor across firms, and a significant increase in productivity losses from misallocation over time. We develop a model of heterogeneous firms facing financial frictions and investment adjustment costs. The model generates cross-sectional and time-series patterns in size, productivity, capital returns, investment, and debt consistent with those observed in production and balance sheet data. We illustrate how the decline in the real interest rate, often attributed to the euro convergence process, leads to a decline in sectoral total factor productivity as capital inflows are misallocated toward firms that have higher net worth but are not necessarily more productive. We conclude by showing that similar trends in dispersion and productivity losses are observed in Italy and Portugal but not in Germany, France, and Norway.

  • El anterior articulo debería leerse conjuntamente con  Political Credit Cycles: The Case of the Eurozone por Jesús Fernández-Villaverde (no precisamente conocido por su simpatía hacia los austriacos), Luis Garicano y Tano Santos. Liga la entrada al Euro con el deterioro institucional del Sur de Europa, de los llamados PIGS. Muy recomendable la parte de España. Su lectura es mucho más fácil que el articulo anterior ya que no requiere conocimiento alguno de álgebra o cálculo estocástico.

We study the mechanisms through which the adoption of the Euro delayed, rather than advanced, economic reforms in the Euro zone periphery and led to the deterioration of important institutions in these countries. We show that the abandonment of the reform process and the institutional deterioration, in turn, not only reduced their growth prospects but also fed back into financial conditions, prolonging the credit boom and delaying the response to the bubble when the speculative nature of the cycle was already evident. We analyze empirically the interrelation between the fi- nancial boom and the reform process in Greece, Spain, Ireland, and Portugal and, by way of contrast, in Germany, a country that did experience a reform process after the creation of the Euro.

Liquidity

Recogemos aquí un artículo de Melchior Palyi sobre el concepto de liquidez. Está disponible en pocos sitios en la red, de modo que tiene su importancia el contar con una copia adicional, para permitir que pueda ser accedido en un futuro.

 

Liquidity
By Melchior Palyi
With an Introduction by Antal E. Fekete

Liquidity was first published by the Minnesota Bankers Association in 1936. The appendix, “Illiquid Central Bank: Graveyard of the Currency” appeared as “Asset Liquidity – A Restatement” in Dr. Palyi’s Bulletin #366, August 11, 1958.

About the Author:

Melchior Palyi was born in Budapest in 1892 and died in Chicago in 1970. He was educated in Hungary, Switzerland and Germany and taught in the Business School of the University of Berlin after receiving his Ph.D. from the University of Munich. In 1928 he became chief economist of the Deutsche Bank, then the largest bank in continental Europe. From 1931 to 1933 Dr. Palyi was adviser to the German Central Bank and managing director of its Institute for Monetary Research. He left Germany after Hitler came to power, and after a short stay in London as the guest of the Midland Bank, came to the United States. He served as a visiting professor and research economist at three universities – Chicago, Northwestern and Wisconsin. Dr. Palyi wrote several books, including The Twilight of Gold: Myths and Realities, The Chicago Credit Market, Managed Money at the Crossroads, and An Inflation Primer. He was a frequent contributor to the Commercial and Financial Chronicle of New York and wrote a weekly financial column for the Chicago Tribune.

Antal E. Fekete, Professor of Mathematics and Statistics, Memorial University of Newfoundland, was graduated from the Lorand Eotvos University in Budapest in 1955 and emigrated to Canada after the Hungarian Revolution of 1956. He studied at the University of Ottawa and Columbia University and has been on the Faculty of Memorial University since 1958. Prof. Fekete has been a student of monetary science and banking since 1959. He is a Research Fellow of the Committee for Monetary Research and Education and is the author of two of the Committee’s monographs: Borrowing Long and Lending Short: Illiquidity and Credit Collapse (1983) and Resumption of Gold Convertibility of the U.S. Currency (1984).

INTRODUCTION

When does a river cease to be a river? At the moment it descends to sea level. Significant and conspicuous changes occur at that point. The ecology of water changes. Water molecules lose their potential and kinetic energy, which is converted into entropy.

Similarly, the flow of myriad goods from producers to market also undergoes a remarkable metamorphosis when it comes within sight of the consumer. Adam Smith noticed this phenomenon when he formulated the concept of social circulating capital. By this term he described the mass of finished or semi-finished goods which has reached sufficient proximity to the consumer so that its destiny of being consumed can no longer be in doubt.

The analogy between the flow of goods to the market and the river emptying into the ocean can be profitably extended to include economic entropy. The risks and uncertainties, so characteristic of production and processing in their early stages, all but disappear by the time the goods form part and parcel of the social circulating capital. Speculation and risk-taking give way to the automatic processes of distribution. Thus entropy can be conceived of as the reduction or disappearance of uncertainty and risk occurring pan passu with the maturation of goods.

The disappearance of uncertainty and risk, along with the emergence of social circulating capital, i.e., the increase in economic entropy, manifests itself in a most dramatic fashion, namely, in the shape and form of liquidity. To Adam Smith, liquidity was tantamount to the spontaneous circulation of real bills in the Manchester and Lancashire of his time. Today, liquidity is a more elusive concept, because of developments in banking and the prevalence of bank loans which have preempted spontaneous bill circulation. Today, liquidity refers to the marketability of bank assets, including the assets of central banks.

Meichior Palyi, an expert on the theory and practice of modern banking, first published his masterpiece Liquidity in the fateful year of 1936, the year John Maynard Keynes’ General Theory of Interest, Employment and Money was published. As a manifesto to stem the incipient tide of “managed money,” Palyi’s work, Liquidity, did not succeed. Latter-day readers, however, may bear witness to the fact that its clear logic, incisive analysis and sound historical perspective are far superior to anything offered in Keynes’ General Theory. To Keynes, “liquidity preference” is the original sin, to be fought with every available means; if necessary, with the strong arm of government. By contrast, to Palyi, liquidity is a pristine virtue, to be protected and preserved at all cost.

In Liquidity, Palyi exposes the inherent fallacy of the quantity theory of money: that velocity of circulation is an independent variable not subject to control by monetary policy. He points out that velocity can be controlled only by a banking policy that respects liquidity. Velocity cannot run away if all bank loans are strictly short term (91 days or less).

Moreover, Palyi explodes the Keynesian edifice of the theory of employment. Banking policy respecting the principles of liquidity promotes employment, whereas monetary policy contemptuous of those principles must ultimatelythwart employment. In acquiring assets, a bank acts as an allocator of capital between long-term and short-term uses. A liquid banking structure tends to give preference to labor-intensive applications, rather than those with larger fixed capital requirements per unit of labor. By the same token, a liquid banking structure strengthens medium-sized business as against the mammoth concern, which is favored by an illiquid system.

Palyi anticipated and refuted the quasi-scientific theorizing of Keynes and hisepigoni, who have argued forcefully and persuasively that it is sound economics and constructive government to finance public works, as well as agricultural and industrial subsidies, through the sale of public securities to the banking system. Palyi notes that such a policy must, in due course, lead to the complete ossification of the assets of the banking system, at which point liquidation of assets no longer is possible except at huge concession in price.

Palyi explains that if the value of bank assets has been decimated and destabilized, then the collapse of the value of the currency cannot lag far behind. There is no way to divorce one from the other, any more than a mirror image can be divorced from its owner. And there is no salvation in central bank intervention. The central bank can only support market values at the cost of making its own position more illiquid. While a reasonably liquid central bank can temporarily “lean against the wind,” an illiquid central bank will be swept away by the whirlwind. The day which sees the bond market seeking refuge in the Central Bank will also see the demise of the currency. Palyi aptly noted that an illiquid central bank is the graveyard of the currency. Therefore, Keynes’General Theory must be seen as a blueprint for the euthanasia not only of the renter but also of the currency.

This edition of Liquidity incorporates minor editorial changes, as well as an appendix: Illiquid Central Bank: Graveyard of the Currency, written by Palyi in 1958. The latter article clearly shows the author’s dismay over the debauchery of the Federal Reserve system in violation of the Federal Reserve Act of 1913.

Without exaggeration, we may say that our comprehension of the passing economic scene cannot be complete without prior understanding of the dichotomy of liquidity and illiquidity concepts conspicuously missing from the vocabulary of most contemporary writers on money, banking and economics. National leaders in government, banking and industry, who desire to reconstruct vibrant national economies and a sound international financial system, would serve themselves well to study the cogent presentation in Melchior Palyi’sLiquidity.

Antal E. Fekete

Memorial University of Newfoundland

St. John’s, Newfoundland, Canada

October 1984

PART I

THE LIQUIDITY DOCTRINE OF LIBERALISM

Eighteenth century writers either took it for granted that commercial banks must be prepared to meet any demand for redemption of their notes by paying out monetary metal, or else they ignored it; but none of them discussed the economic implications of a short-term credit structure. Since Adam Smith and publication of The Wealth of Nations (1776), short-term credit structure has become an integral part of the system run by economic man in accordance with his rational self-interest and long-term point of view.

According to the theory, as developed by H. Thornton (1802) and J. Fullarton (1844), banks do not necessarily add to the volume of circulating media, but only “monetize” such credit instruments as have existed before into a more readily circulating form. By the very nature of their business, banks can only temporarily raise the volume of money; the backflow of their automatically self-liquidating, short-term credits limits both the size and the duration of the expansion. The banking mechanism is such as to adapt the credit volume to the flow of goods in an “elastic” fashion.

This theory, developed fully in the first half of the nineteenth century, soon absorbed three major modifications. The first had to do with central banking and was the outcome of the lengthy quarrel between the Banking School and the Currency School: it was asserted that things do not work quite so automatically, and therefore the central bank must apply the brakes to avert overspeculation and to moderate panic. Secondly, it had to be recognized that the rules for liquidity of bank loans do not always, or at least not fully, apply to banks’ (secondary) reserves, for which marketable securities, especially Treasury bills, may offer a more readily liquidated form of investment. Of course, this was not supposed to amount to more than a moderate portion of banking resources.

A third modification developed in Germany. Germany banks, from the beginning, violated the classical rules of liquidity. They combined commercial banking with investment banking, and financed industrial development on a rather nominally short-term basis. Accordingly, German experts argued that the basis of liquid credit had been broadened by including goods in the process of production, in addition to those in the process of commercial transactions. The German Kontokorrent-Kredit has been invested, by this theory, with the attributes of the English commercial bill of exchange; the liquidity concept has been stretched to include working capital provision as well. Yet the classical theory retained its predominant position, not only as a postulate of what banking ought to be, but also as an alleged description of what it is. This in spite of the fact that commercial banks (with the possible exception of a few leading institutions in France and Holland) have already been deeply involved in the securities business and long-term finance. The Germans approved or at least faced this development, while the British deplored or tried to overlook it.

PART II.

A REVOLUTION IN MONETARY THOUGHT

Not until World War I had shaken mankind’s belief in “fundamentals” did a general attack on the liquidity ideal itself arise. In America, it came especially from the financial frontier in the Midwest where heterodoxy in such matters had long been popular, and where even the most vehement cyclical convulsions failed to shake the optimism of a community of speculative pioneers in history’s greatest real estate development.

Teaching and banking traditions have accepted, in America as elsewhere, the British belief in liquidity; but banking practice has been far from accord with it. Consequently, it offered an easy target for the attack of the Institutional School. In 1918, in a series of articles in the Chicago Journal of Political Economy, H. G. Moulton carried out one of the most brilliant attacks of this school against the traditional doctrine. His point was that very few commercial loans could be relied upon for liquidation; and in time of crisis no liquidation is possible at all. The loans are based much more on a permanent “alliance” of the banks with other business units than on the financing of specific completed transactions, and are less “liquid” than marketable securities. Liquidity, indeed, in the sense of liquidation, has meaning only for the individual firm. The banking system as a whole, Moulton thought, does not know any liquidation other than the shifting of assets from one bank to another. Shiftability takes the place of liquidity; banking becomes a matter of choosing the properly marketable assets, and banking policy a matter of securing mechanisms to create or maintain shiftability.

This doctrine, conceived at a time of prosperity and credit expansion, seemed in perfect conformity with modern development. It gives a quasi-scientific basis to the old request that commercial banks furnish industry with at least its “permanent working capital,” and give up the idea of purely short-term credit-leading even such cautious documents as the British MacMillan Report of 1931 to the proposal of “more closely coordinating” the financial organization of the City with British large scale industry. It is also the basis upon which the propaganda for public works and subsidies to be financed by the sale of public securities to the banks, has been built up during this depression. Furthermore, it underlies the new monetary ideas so important for our period-ideas of monetary control based on some sort of quantity theory approach.

At the beginning of the twentieth century, the tradition of the quantity theory of money had been represented by only a few mathematically-minded economists and by money “cranks.” The monetary and banking systems seemed so solidly based, and so little exposed to outside interference that the quantitative approach had only a purely academic interest. The breakdown of leading currencies during World War I taught a new experience, and showed that monetary control might be used for almost any purpose. It coincided, and not merely by chance, with the rediscovery (also by American students, Davenport and C. A. Phillips, 1916) of the fact that a major part of deposits had been actually created by the banks themselves. The Moulton School, however, provided the basis for use of the quantity- theory type approach for policy purposes. If credit lacks “quality” except of some artificial and readily creatable type, then, of course, the purely quantitative manipulation of the credit volume is the “real thing.”

Practically all currency reformers, aiming at some sort of price, or income or employment stabilization by the control of monetary volume, have their common foundation in Moulton’s criticism of the traditional liquidity principle. Consequently, the term “liquidity” does not even occur any more in most current books dealing with the theory of banking, or with the business cycle. Their interest centers on “measurable” quantitative problems, and the control to be exerted over the volume of money. “Qualitative” problems of bank policy are either ignored or ridiculed.

Contrary to the communis opinio of previous generations, most monetary reformers deny the relevance, or the very existence, of any fundamental distinction between short-term and long-term investment, between a bill and a bond, a note and a mortgage. They deny common sense opinion according to which the quality of bank assets is largely responsible for cyclical fluctuations. The banking apparatus is supposed to be able to generate or to destroy credit, to any extent and arbitrarily, depending solely on legal or traditional cash-reserve requirements.

On the other hand, most of the “old timers” still like to argue against the simple arithmetics of the quantity theory and to overemphasize the qualitative aspects of bank credit. They like to assume that banks do not exert any control at all over the volume of credit, and argue either for freedom of commercial banking, or for interference limited to mild rules of liquidity (eligibility). It is hard to say which of the two schools of thought is less realistic. They are both guilty of ignoring problems which they do not seem able to incorporate into their line of thought.

PART III.

THE MEANING OF LIQUIDITY

The prevailing confusion is largely due to thoughtlessness in the use of the term “liquidity.” It is often confused with a concept of physical type: working, as opposed to permanent capital. But the concept is meaningless without reference to contractual obligations.

Liquidity, at first sight, is the capacity to fulfill financial obligations. This, in turn, is not identical with cash (prime) reserves. The cash ratio is a minor issue compared with the status of the bank’s earning assets. If these are “liquid”, the necessary cash, given a customary minimum, is easily found. The gradual decline in England of the ratio of cash to sight liabilities from 25 per cent to 30 per cent in the eighteenth century to about 6 per cent to 7 per cent in the 1920’s (the latter consisting mainly of balances with the central bank) is by no means unsound in itself. It must be viewed with due regard to changes in the structure of liabilities (deposits instead of notes); to growth in the use of money substitutes (e.g., checks); and to changes in the composition of the banks’ earning assets. Less natural since the end of World War I was the reduction in the reserve ratio in the American city banks from 25 per cent to 10 per cent to 13 per cent for checking deposits and to three percent for savings deposits; and the practice of German banks of keeping cash holdings down to some two to four percent. Most questionable, at any rate, were the practices in America to include interest- bearing balances of other banks among cash items, or the English routine to count those credits at call or on short notice as “till money.”

In reality, the long-term trend of reduced cash holdings is not due to the improved liquidity of earning assets, but rather to market developments permitting the sale (shifting) of assets on a large scale.

Stock markets have developed to unforeseen extents; central banks and even governments have put their resources at the disposal of banks so as to make liquidations possible, etc. These trends point to the “relativity” of the liquidity concept. The standards of both, the cash ratio and the liquidity of earning assets, are determined by a bewildering number of factors. They will depend, for example, on such facts as the confidence of the public in the banks. Optimism or pessimism of cyclical character are even more important. Established standards of what is proper practice exert a great deal of “irrational” influence, too. Still more important is the general monetary organization of the country. Liquidity of banks is an entirely meaningless concept in a progressive currency inflation, the ideal of which is to escape the impending depreciation of liquid funds. (In 1923, the Germans called itSubstanzwerte, meaning everything from undeveloped real estate to empty matchboxes.) A currency unit with widely fluctuating gold content allows the banks to compromise substantially the standards of credit discrimination; the very term “liquidity” is tied up with a currency system which limits the amount of available cash according to the “rules of the game.”

Most of the confusion arises, however, from the fact that liquidity is generally thought of as the ability or readiness to “liquidate.” The shiftability approach argues that there is no liquidity at all, since the whole system could not be liquidated, and overlooks the possibility or danger of some partial liquidation. The problem of one bank might be successfully eliminated if the others are willing and sufficiently liquid to take care of it; or if the current growth of savings covers the bank’s deficit and if it flows in the desirable direction; or if the government steps in; or if foreign help is available. Perhaps some combination of all “shiftings” may do the trick and postpone the evil day. But there is no use trying to eliminate the problem by wishful thinking, which ignores the fact that the total of the banks’ assets cannot p05sibly have a book value greater than the total of their liabilities. Consequently, bank deposits should at all times be capable of buying the assets. Whether the owners of those deposits are willing to buy the banks’ assets raises the question of prices. The demand for such a large variety of goods as the assets of a national banking structure can hardly ever be altogether inelastic. It may cost terrific price cuts to sell out, but it is useless to argue that there is no problem of liquidation because assets could not be sold out wholesale. The argument ignores the possibility of liquidation at falling prices. Even land might be liquidated en masse, as most of the real estate in Berlin changed hands during the 1923 inflation when prices, in gold, fell sufficiently.

However, observance of liquidity rules does not imply preparation for liquidation. On the contrary, liquidity means preparation-for the avoidance of liquidation. The periodic liquidation of each individual or short-term bank transaction should not be confused with the liquidation of any part of the total.A liquid structure never liquidates; only the illiquid one comes under the pressure of liquidation. ‘Perfect liquidity” means that, for any length of time, all financial obligations are fulfilled without net liquidation of capital. A liquid society has adjusted its obligations to the flow of its income, both in amounts and in maturity dates, so that forced sales should not occur (disregarding war, or other extra-economic factors). An open illiquidity (as opposed to a concealed illiquidity) means either a refusal to pay (i.e., collective bankruptcies, moratoria and foreign exchange controls), or the necessity of forced sales of bank assets, or both. The former method eliminates the problem by uprooting the legal and credit structure; the latter restores liquidity, but at the expense of crises and depressions.

PART IV.

THE BURDEN OF ILLIQUIDITY

For an enterprise which “lives” on credit-making, the issue of liquidity virtually coincides with that of its earning power. The bank’s earning power depends on the “credit” of the bank which is based on the assumption of its liquidity; and this assumption in turn vanishes if the bank ceases to be a going concern. Now earning power, in the first place, is a matter of costs. Their rise typically foreshadows growing illiquidity. For banks, more than any other line of business, long-term earning power is a matter of provision for losses. Bank liquidity, therefore, begins with an adequate capital ratio (i.e., the ratio of properly invested net worth to liabilities). The conspicuous decline of this ratio in the balance sheet of commercial banks during the past century is due to causes similar to those of the cash ratio: from 1:3 to about 1:8 in the United States, and to something like 1:13 in England and France, and to even less in Germany. It was about 1:25 in the Danatbank, the failure of which in 1931, losing more than its capital in a single credit transaction, precipitated the Berlin crash. But from the liquidity angle, both the net worth and the cash reserve are only minor considerations. Both represent the immediate or tactical point of view, rather than the far-sighted or strategic one, that of the liquidity of earning assets.

In this respect the first choice is between short-term loans and investments (bonds). The latter are allegedly far “safer.” But, between 1902 and 1914, for instance, in a period of balanced budgets, one of the English “Big Five” banks had very severe losses on its unusually large holdings of British consols (perpetual bonds) which have been considered the most “solid” and “stable” investment of the world for almost a century, but then fell, as many times before, with the upward

trend of the cycle. Even “first class” long-term paper involves very substantial risks, due to fluctuating market quotations. To avoid losses, banks are compelled to sell out holdings of securities whenever their prices fall continuously; this is a typical case of a perfectly “good” investment which causes liquidation and therefore has not been “liquid.” If over 50 percent of the assets of American banks are now invested in government bonds (most of it of a long-term type), the dangers are serious indeed. A much more pronounced condition obtains in Germany and Italy; a less serious one in England. It is not as if the breakdown of public credit would be an imminent danger, but a minor fall in the prices of those securities wipes out the earnings and even the capital of the banks-to say nothing of the danger to the value of the non-marketable long-term claims of the banks. It goes without saying that only “shiftable” paper is advisable for either the secondary reserve or the investment portfolio of commercial banks. The practice of many American banks to invest major amounts in mortgages was exceedingly dangerous, especially when it was done on the basis of reckless overvaluation and almost criminal disregard for the elementary rules of prudence.

The difference in maturities means a great deal more than the heavy risk of fluctuating values. The longer the duration of the loan, the more knowledge about future conditions is needed for the proper assessment of credit. This raises the question of capital loans to industry. Interweaving credit-granting with commercial transactions permits an insight into their nature, and thereby into the risks involved, which has to be substituted otherwise by an intimate knowledge of the whole business and its prospects on a much wider range. True, in Central Europe, there is a type of versatile banker who is supposed to handle the problems involved in industrial finance as much as the old-time banker handled commercial bills. But the results are such that one is having doubts about the social value of the financial superman, to say nothing of the advisability (and possibility) of breeding him in larger numbers. Even adherents of the shiftability theory are increasingly inclined to recognize that it is in the sphere of security and mortgage investments, and long-term industrial credits, that by far most of the banks’ mistakes and losses occur.

Furthermore, short-term credits imply automatic backflow which means very little if, for example, American industrial customers liquidate once a year and have their credit restored a fortnight later. The principle of reflux, if properly applied, helps to control credit in two directions: the total volume expanded, as well as its use for short and long-term purposes. It is a somewhat mechanistic but very useful device to supplement the bankers’ judgment of the credit risk-or to check on it. This check is missing, ex definitione, in the case of long-term credits.

The greatest risk, however, in credits which provide working or fixed capital, is the threat of their permanent renewal and expansion. The underlying assumption of such capital provision is that the high profits of the debtor, and a flourishing capital market, will take care of the bank credit in due course. This forecast may be borne out in good times. But a banking structure which embarks on large-scale financing in advance of future security issues runs even more risks than the excessive danger of immobilization of bank funds. Good money may have to be thrown after bad, in order to forestall the total loss of the original investment. The interconnection of industry and finance due to this combination of commercial and investment banking means the control of banks by industry more often than it means the opposite.

According to most current standards, the bank has done its duty when it has used its surplus funds for “proper” collateral loans. This policy, so far as it goes, safeguards the banks from losses. In fact, the banks rarely sustain losses on stock exchange loans. During the last crisis, credits to speculators turned out, to all appearances, the “safest” way to entrust the depositors’ money! And the experience of previous crises with many lombard loans on paper or on goods with plenty of “margin” has been similar. There was, however, the proverbial “fly in the ointment.” The banks had to liquidate the same kinds of collateral which they themselves owned, and endangered the solvency of their commercial customers by forcing sales on the part of collateral debtors.

Credit on collateral is perhaps the most crucial problem of bank liquidity. Used as a technical term, it simply means additional safeguards for the loan, without any implication as to its purpose. In the economic sense, it is distinctly different from a commercial loan because it generally is divorced from any genuine transaction in the course of the “normal” sale of goods. The problem is especially relevant in view of the fact that collateral loans are likely to be the first line of defense in case of a drain on the bank’s cash resources. They may readily be turned into cash and, therefore, “liquid” from the point of view of the individual bank. But, for the banking system as a whole, collateral loans in great amounts represent the most serious danger of illiquidity. They involve the necessity of liquidation which in a crisis may save the single institution, but only at the expense of wholesale liquidation with its deflationary consequences. The disastrous effects of the huge amount of lombard loans on the Paris stock exchange in 1857, or of brokers’ loans in New York, especially in 1929, etc., are generally known. The latter were particularly disastrous, since the eight billion dollars in question represented largely the “liquid” reserves of the provincial banks.

The technique of deposit creation through bank-to-bank credits is another aspect of the same principle. The process is typical for almost every period of “prosperity.” Finance bills were the instrument by which the most notorious speculative ventures had been countenanced, ending in disaster. Baring Brothers of London failed in 1892 with a ratio of 1:4 between capital and acceptances. A more unfavorable ratio was again characteristic for many London acceptance houses by 1929. The quantitative expansion of credit is especially important when it indicates a deterioration of quality. Before World War I, the balance sheets of German banks showed for a long time a more rapid growth of acceptances than of deposits, and the German experts became suspicious of this inflationary practice by which the competing banks diverted money market funds to their industrial customers. The jittery 1920’s revived this age-old technique of prosperity-makers. It began with legitimate acceptances with shipping and insurance documents attached. Gradually, the documents were dropped and eventually every reference to the commercial transaction disappeared. Next, the “quasi-reimbursement” was replaced by simple bank-to-bank credits transferred on the wire in fantastic proportions. The creditor banks had helped to finance the boom by “confining” themselves to a most “liquid” asset, to credits granted to other (foreign) first class banks. What could look more liquid than a balance with an A-1 bank? But what guarantee did the creditor have that the debtor bank would keep liquid in its turn?

PART V.

BANK ASSETS AND THE MONEY SUPPLY

Let us assume that the banking system is granting credits solely of the short-term commercial and clearly seasonal character, and is being managed so as to avoid major mistakes. At given prices, goods would be sold and debts to banks repaid seasonally not by shifting them to other banks, but by using the deposits of the purchasers. As seasons do not coincide in all trades, some firms take fresh credits at the same time as others retire old ones. Seasonal fluctuations not ironed out automatically could be taken care of by an active Reserve Board. At falling prices, substantial credit margins having been assumed, the banks cannot suffer losses. There is no reason why such a system should get into liquidation on “endogenous” grounds. Nothing in its own structure could cause liquidation and runs are not likely to occur since mistrust in a system which is not “frozen into any major loss-generating venture is hardly possible. Provided that the banks are properly managed, their funds have been used exclusively for such ventures in which the danger of unsalability of goods (within reasonable time) is practically excluded; “speculation,” by assumption, has not been financed with these funds.

Now let us make the assumption more realistic. Suppose the country enjoys a balanced public budget, and the government has a seasonal demand for short-term funds which would be properly satisfied by commercial banks. Similarly, the banks can, with proper caution, engage in short-term operations in foreign money markets. Of course, a reasonable cash reserve, and an amount of “secondary” reserve in the form of first class marketable securities may be taken for granted, too, the latter corresponding on the whole to the genuinely long-term funds at the bank’s disposal.

Obviously, the previous conclusions still hold true under these more relaxed assumptions. The decisive point is that the volume of normal commercial transactions, disregarding seasonal fluctuations, is hardly ever subjected to violent changes. Speculative activities and the flow of savings into investments may dry up, but the basic commercial life which provides the consumers’ current needs cannot stop. Nor are banks ever reluctant to finance it. The English MacMillan Report of 1931, as well as the Hardy-Viner Report to the American Treasury (1934), both keenly desirous of reform, reaffirmed the old experience that strictly commercial credits are always available in a modern banking community, and at a reasonable rate of interest. Nor under the conditions described, need changes in technology or consumption have major liquidating effect on the banks’ total credit. They would cause only permanent shifts in the distribution of credit among debtors, just as seasonal fluctuations change it temporarily.

Of course, “extraneous” factors, such as international conflicts and revolutionary changes in the legal basis of social economy, may still upset the stability of this order. It does not imply a panacea against minor fluctuations, either. Assuming that a breakdown might still occur, the very fact of bank’s liquidity would have the effect of reducing the impact of a depression. The banks would not get into trouble – by definition. They would not incur losses, and would not suffer from panicky fear of the public. Nor could they be forced into major liquidation. So long as the credit they granted has been of genuinely short-term character, their “automatic” repayment would not be in danger. If, with decreasing trade, the volume of fresh credits should be reduced, this deflationary process would be very mild compared with the usual one in a crisis, because it does not involve the necessity of forced sales on any scale similar to that experienced under conditions of illiquidity. As a matter of fact, the intensity and length of the crisis depend largely on the resistance which the banking structure is or is not able to offer. An illiquid structure leads to a crash which a liquid one not only avoids for itself, but may actually soften for the rest of the community, by being able to “come to the rescue.”

The main point, however, is that if bank credit is provided largely on short-term commercial lines, its total volume cannot exceed the demand for circulating capital proper, i.e., a sum commensurate with the amount of goods flowing to the market at prices at which they can be sold. As a matter of fact, bank credit should lag far behind this amount, because not all such transactions need to be financed by banks, not all who may need financing are sufficiently good risks, and all commercial goods should be financed only with a substantial margin. At any rate, the total volume of circulating media is effectively limited by the observance of liquidity rules. It is limited, as D. H. Robertson has pointed out, to a level far below the amount of dollars which represent the value of the circulating capital of the country. Tile discriminatory choice of bank assets amounts to a restriction of the volume of deposits, within narrow limits. Of course, during the short period between borrowing and repayment, the borrower draws on his balance to make payments which in turn may swell the deposits of others. But this credit expansion is, so long as the banks adhere to the rules of liquidity, under a two-fold quantitative control: its volume is comparativelystable, since violent fluctuations and forced liquidations are not likely to occur; and the total amount is limited by the short-term commercial credit demand and cannot be extended far beyond it. There is, under the assumed conditions, no “automatic” expansion to the limits permitted by the cash reserves.

It is misleading, however, to assume that the bank’s liquidity is identical eo ipsowith a stable and entirely undisturbed money supply or price level. But the disturbances in question are, by the nature of the system, greatly reduced in comparison with an illiquid structure. The liquid structure limits the possibility of fluctuations by not allowing the banking machine to supply more currency than is compatible with the volume of goods forthcoming, within a short time, at given prices. And liquid banking makes it possible to exert influence by discount policy on the demand for bank loans which proves “inelastic” under other conditions. A money market which serves largely long-term investment purposes is hardly capable of adapting its credit volume to changes in the rate of interest. The classical theory of money-market control by discount-rate changes and by open market operations was based on the assumption of a liquid banking structure. A liquid banking structure allows the central bank or the Federal Reserve system a substantial power over market fluctuations. The actual failure or unsatisfactory working of discount and open market policy in major booms and depressions reflects the fact that the banking system has been illiquid in each case.

And this is not the whole story. Theoretically, a quantitative policy can be devised to “manage” the money supply according to preconceived standards. But monetary management per se must turn out to be a failure if the banks have already committed themselves along illiquid lines. Interference then leads to breakdown, which it was supposed to avoid. Liquid banking, on the other hand, actually achieves “stabilization” by inhibiting the major boom and eliminating its ‘consequence, the major depression. In addition, pure monetary control is limited by the difficulty to control money’s velocity of circulation. Velocity is known as an independent variable of the Equation of Exchange. It does not necessarily vary directly with changes in the money supply; it may vary inversely with it. Consequently, control over the money supply in itself is not sufficient to control price or income level fluctuations, since changes in velocity are usually beyond control. Liquidity policy, on the other hand, has the advantage of indirect control over velocity, too. The shorter the period between the lending of funds and the repayment date, the less the likelihood of repeated use of the deposits. The number of times a deposit can be use for payment is naturally limited by its lifetime, which depends on the duration of credit for which the deposit was created. Furthermore, liquidity means qualitative credit control checking the speculative activities of the boom which tend to increase the velocity of circulation. It also counteracts hoarding tendencies during depression, thanks to the stable volume of commercial credit, the avoidance of forced liquidation on the part of banks and the elimination of runs on them.

PART VI.

CAPITAL ALLOCATION AND CREDIT POLICY

The banking system not only creates means of payment, but also allocates them. The purely quantitative approach does not bother about the second function, which is, however, not of minor social importance. Everybody knows that banks use their lending power in a discriminating fashion. But it is not common knowledge that the character of this discrimination regulates the effective volume of currency. Nor are the allocating effects of the process generally appreciated.

The choice of banks’ assets is a directing factor in the allocation of capital between long-and short-term uses. A liquid structure tends to give preference to “labor intensive” industry, as against the one with larger fixed capital requirements per unit of labor, and ceteris paribus, to a commercial enterprise rather than to an industrial one. The preference for providing circulating capital also tends to strengthen the medium-sized business as against the mammoth concern which in turn is favored by an illiquid system. Of course, banks are only a minor force in determining the industrial structure, but they can contribute to it in a significant way. The shifting of bank funds into long-term industrial finance attracts, in that direction, other funds as well. The industrial development of countries has been deeply influenced by such practices which aided the growth of large-scale units far beyond the point of optimum size. This does not mean, however, that liquid banking protects the small unit against technological progress, or the “established way of doing things in the face of the competition of newer ways”, as C.O. Hardy said. The industrial and commercial unit of the horse-and-buggy age has practically no access to the lending counters of well-managed banks, while illiquid bank credit has helped many inefficient units to survive longer than socially desirable.

It is no mere accident that countries in which banks are continuously engaged in long-term industrial finance (as in Italy and Germany), or in the financing of industrial securities (as in the United States), have witnessed a most spectacular growth of large-scale units and monopolies. In England, on the other hand, and especially in Holland and France, where liquidity rules were abandoned at a less rapid rate, the development of large-scale units and monopolies was much slower and the independent units, both in manufacturing and in wholesale trade, had a much better chance for survival. This difference had nothing to do, apparently, with branch banking; bank concentration in England and even in France has progressed virtually as far as anywhere else.

It is difficult, however, to estimate the exact extent to which banking policy influences such long-run developments. The cyclical influence is more easily appreciated. Two points must be emphasized in this connection. First, the fact that a banking system’s choice of illiquid assets works itself out in a cumulative way. Suppose, for instance, it buys mortgages on a large scale. At first, the marketability and value of mortgages will tend to rise and, consequently, new borrowers would have even better opportunities to obtain more credits on similar assets. The credit policy of the banks influences the allocation of capital far beyond the volume of bank resources deployed. Secondly, the intensity of “speculative” activity is largely a matter of distribution of loans (advances) and investments by the banks. The flow of bank funds into specific channels may start or accelerate the rhythm of speculation along those lines, by generating psychological forces so characteristic of aggressive business optimism. Liquid banking, on the other hand, implies control over the use made of borrowed funds and the probability that they will be applied to “productive” purposes. The opposite policy opens the door for indiscriminate financing of all sorts of ventures which so often turn out as “bubbles” or other malallocations of capital.

Wasteful speculative orgies and malallocation of resources with ensuing losses cannot, of course, be entirely eliminated, and may come about without any banking support. But it is most important that they should not be magnified into catastrophic dimensions. This depends mainly on the banks’ policy in choosing their assets. Indirectly, by sustaining a credit inflation, and directly, by financing maldirection of capital, the banks carry the responsibility for disaster. The growth of the economic system may be such as to offset, by fresh “real” savings, the capital losses due to “unproductive” investments. However fictitious the assumption of such growth may be, it typically underlies the philosophy of periods of prosperity.

Although the phenomena of the business cycle is commonly formulated in terms of a disequilibrium between the effective money supply and the flow of goods, or between the flow of savings and the volume of investments, etc., such quantitative formulas tend to overlook the fundamental chain of causation. It is the wholesale financing of abortive ventures with the aid of bank credit expansion which generates the boom. And it is the breakdown of these ventures and the sudden drying-up of the flow of bank credit which necessarily brings the boom to a halt. The purely quantitative approach neglects this allocative effect of the banking process. It does so by throwing overboard the principles of liquidity. But what other reasonably practical standards for limiting the volume of currency can be substituted? No two monetary reformers agree on what measure of the money supply should be stabilized; nor on the technique by which to achieve it. And no “stabilization” can overcome the difficulty that, whatever purely quantitative levels one chooses, they will either inhibit legitimate growth, or else permit illegitimate over-expansion.

Ultimately, the choice is among three possible lines of policy: the old-time ideal of laissez-faire, which leaves banks free to follow the vagaries of business psychology; the new religion of “controlling the money supply,” handing all power over the credit structure to political forces; and a policy of cooperation between a liquid banking structure and an active Federal Reserve. The first is hardly worth discussing, in view of the violent fluctuations of trade which it implies. The second promises stabilization, but has no way of eliminating the danger of illiquidity. There is no escape from the problem of liquidity; it is identical with that of right or wrong investment. The very meaning of banking as a social function is to supervise the channels into which the flow of capital is directed. It exerts this function by using the “liquid” funds of society in a way which, so far as humanly possible, avoids losses and forced liquidations. Monetary control believers either ignore this aspect or else assume arbitrarily that mere manipulation of credit volume will somehow solve the issue.

A credit policy that neglects liquidity standards has the great advantage of permitting-in theory-“eternal” low interest rates and the development of “new eras” of apparently limitless expansion. So long as depression prevails the dangers involved are not likely to impress limited imaginations. With the change in the cyclical outlook, however, the problem will reappear soon enough-unless dynamic factors, such as population growth, technological progress, and speculative enterprising, should be virtually eliminated. As soon as major speculative activities develop, even the most intelligent monetary management (and who dares to assume that it will always be intelligent?) cannot do much by relying solely on quantitative standards. The wildest sort of speculation was characteristic of the 1920’s without any major rise in the general price level. In America, while the gambling orgy was most intense in 1928 and 1929, the volume of demand deposits subject to check hardly grew at all. It was the deterioration in the quality of investments,.as measured by the illiquidity of bank assets, which engineered the liquidation as soon as losses became visible. The policy of “stabilization” of the Roosevelt Administration can, of course, be carried further by credit inflation and devaluations. The depression may be avoided (or more precisely, its impact may be reduced) at the expense of the currency and its stability. But only very strong countries can afford such a drastic cure more than once; and it is very doubtful indeed whether the decline of international trade, the tariff warfare, and other worldwide economic and political repercussions are not too high a price to pay for the temporary enjoyment of a boom.

Liquidity policy, on the other hand, does not rely solely on the qualitative control of bank assets. Its standards also imply, as has been pointed out, the control of the volume of circulating media. Moreover, it implies such control in advance,before the unsound development has taken place; not as the quantitative control does, postfactum, when it is too late. But to be effective, it must be supported by active Federal Reserve policy. As an efficient institution, driving to stabilize the foreign exchanges and to straighten out major internal fluctuations, the central bank is the correlate to a liquid commercial banking structure. They mutually reinforce one another. The central bank’s function is to set and enforce liquidity standards. Its “moral” and other powers, supported by legal requirements if necessary, go a long way. The belief in “free banking” implies more grave errors than one. It assumes far-sighted wisdom on the part of all bankers. It assumes that enlightened self-interest is a simple rule. It overlooks the fact that the banks are mostly, by their very nature, under the influence of external forces, especially of monetary policy (or lack of it), and of business psychology. The choice is not between “free” and “regulated” banking, but between right and wrong leadership.

Liquidity of banks is a “limiting” case, or an ideal. For practical purposes, what matters is the degree of actual approximation. The strongest argument against this ideal is derived from experience, which apparently shows that liquidity standards are invariably waived in periods of over-confidence. But closer scrutiny of such experience would undoubtedly show the responsibility, for a large part, of governmental and central banking policies. They are responsible, at least in a negative way, by having neglected to use their powers in due time to enforce liquidity standards.

Fortunately, tradition and self-interest of the financial community tend in the direction prescribed by the ideal. Its enforcement is therefore more a matter of maintaining traditional standards than of using “force.” This points to other fundamental differences between liquidity policy and a purely quantitative regulation. The first means active cooperation between central bank and bankers. It leaves the latter to carry their full share of responsibility, but it helps them to understand and to maintain the proper standards. It also presupposes public financial policies (balanced budgets) which do not compel the banks to buy government paper. The other, quantitative regulation, throws the entire responsibility for success or failure on the central bank or the treasury. It leaves the banks free, or actually encourages them to finance whatever abortive ventures capture their fancy and it permits the government to embark on deficit financing. The one strengthens the natural interest of bankers, businessmen and authorities in sound financial standards and tends to eliminate such leadership which is not able to live up to them. The other tends to “institutionalize” unsound financing by eliminating its strongest institutional and psychological hindrance: liquidity. In last resort, the one policy is part of the liberal ideal which thinks of economic restraint of the individual as a social necessity and of failure as a “just” punishment for violating the rules of the game. The other is in conformity with the unsound idea which knows no difference between “legitimate” capitalist business activity and sheer speculative gambling, which tends to concentrate all economic power in the hands of centralized autocratic bodies, and which tends to substitute for the idea of competitive fairness, the ideal of safeguarding vested interests, at whatever social costs.

The essence of liquidity is maintaining the currency in a readily moving” condition, so as to avoid its freezing, and the ensuing “stickiness” of prices. It is, therefore, a prerequisite for the maintenance of the gold standard which implies liquid banking as part of its rules. And liquidity standards are fundamental to any policy attempting to keep the economic system in a state which enables it to cope with a changing world without being uprooted.

APPENDIX

ILLIQUID CENTRAL BANK: GRAVEYARD OF THE CURRENCY

Liquidity of the banks’ earning assets is as important for the solvency of the individual institution as it is for the stability of the economic system. The social objective of banking is to furnish the liquid funds necessary to keep the economy in operation and expansion-without inflating or mal-allocating them to the extent of bringing about boom-and-bust cycles and monetary crises. Asset liquidity prevents their occurrence. By contrast, “managed money” attempts tocure them after the event.

The classical pattern of asset liquidity, formulated by Adam Smith (1776) is known today as the “real bill” concept of asset liquidity. Accordingly, short-term commercial paper, or its equivalent, representing the actual sale of commodities, constitutes the proper realm of commercial bank credit. No “over-issue” of currency or deposits can occur as long as the banks finance strictly self-liquidating short- term transactions. The credit operation is consummatedpari passu with the merchandise “change of hands”; no imbalance between the supply of money in circulation (aggregate demand) and the supply of marketable goods has been created. If banks confine the use of their liabilities subject to quick withdrawal to such self-liquidating assets, the purchasing power they generate would be limited to the value of goods in process of marketing or production, at current prices. By the same token, any addition to the amount of circulating media arising out of the direct and indirect financing of long-term or not self-liquidating ventures risks unbalancing the overall demand-supply situation and “immobilizing” the credit institutions.

The theory is borne out by more than four centuries of experience with business cycles. Witness the history of modern crises, reaching back to the recurrent waves of Venetian bank failures in the early sixteenth century. In every instance, the wholesale liquidation of debts was the focal point, brought about by a credit expansion along non-commercial lines, financing long-term loans, speculative ventures, and governmental expenditures on a substantial scale.

The chief departures from the classical principle of bank liquidity are three.

Since the turn of the century, Germany experts claimed that banks could provide business with working capital rather than with circulating capital – a subtle distinction. They rationalized the widespread practice of Continental institutions which used to finance unsold and often unsalable inventories, despite the fact that time and again the credits turned out to be “frozen.” Substituting money market funds for those of the capital market is typical of new industrial countries in rapid growth and short of capital. It results in a constant reliance of the commercial banks on the central bank, in recurrent bank failures, and in severe cyclical repercussions.

Another-purely academic-school of thought negates the concept altogether, arguing that the banking system as a whole could not hold its assets in a form fit for liquidation. Therefore, supposedly, it is futile to attempt to maintain a liquid status beyond the cash reserves needed as a matter of routine. Runs on banks prove, allegedly, that nothing short of 100% cash liquidity could stop them. In reality, runs prove the exact opposite. They do not even occur, barring extraordinary circumstances such as a major war, unless the banks are known or believed to be in an illiquid condition.

Briefly, the rationale of maintaining asset liquidity is to avoid the occurrence of conditions which may bring about the wholesale liquidation of debts. Though “perfect” liquidity cannot be attained (no more so than “perfect” competition), its approximation is a first essential for all banking responsible for carrying the cash reserves of their customers, of the nation.

Still another school contends that, as a matter of fact, the banker is interested in shiftability rather than in liquidity. Shiftability means the ready marketability of assets without loss, and puts the emphasis on the collateral behind the loan rather than on the nature of the underlying transaction. This presupposes, in effect, security markets at stable quotations. But then, what guarantee is offered for the continued availability of such outlets? The shiftability concept, as interpreted in the 1920’s, assumed that a buoyant stock market, capable of absorbing new security issues ad libitum, would enable the corporations to liquidate their bank debts. The Great Depression thoroughly deflated this theory-and opened the door to a new version of it, legalized in the Banking Act of 1935 that made “sound assets” eligible for rediscount at the Federal Reserve Banks.

Since World War II, the problem of shiftability has been “solved” along the lines followed by the German and other Continental banks after World War I. Liquidity of earning assets became virtually synonymous with their “rediscountability” (in the vocabulary of Federal Reserve Governor Marriner Eccies). Regress on the money-creating potential of the central institution provides the “market.” AsAPlan for Member Bank Reserve Requirements of the Economic Policy Commission of the American Bankers Association stated in 1957:

It is now universally recognized that for the banking system as a whole, liquidity depends, ultimately, on the ability and willingness of the Federal Reserve to supply additional funds to the banking system in periods of stress.

It is not clear what is meant by the system as a whole-as different from the individual banks that constitute it-unless it refers to a money market generally under “strains and stresses.” In any case, the idea that the central bank is to serve not only as the “lender of last resort,” but also as the guarantor of the credit structure’s liquidity, is by no means universally recognized. However, it is being almost universally practiced. The Federal Reserve system fulfills this function by using its resources to provide a market for certain types of debt certificates issued by the federal government.

Government debt certificates-overwhelmingly short-term obligations-are virtually the sole components of the Federal Reserve System’s non-gold holdings. Consequently, those debt certificates have become for all practical purposes equivalent to cash; as such, they constitute the “quick assets” of the commercial and savings banks. They are liquid because they can be readily monetized at the central bank. But their volume is totally divorced from the economic process. It depends on whether or not the Treasury runs a cash deficit, on its debt management policies, on its propensity to take recourse to the facilities of the banking system, and on the latter’s readiness to “oblige.”

Nominally, the Federal Reserve System, at its discretion, may monetize or may retrace its steps, expand and contract. Thereby, the automatism of the commercial credit setup under a self-regulating gold standard is replaced by central bank authority in charge of managed money. This tremendous power implies responsibility-to whom? Needless to say, the ultimate power belongs to those who had delegated it, be it the Congress or the Administration. As a matter of fact, the Federal Reserve may be in actual control of the money supply while things go smoothly – as long as the government finds no difficulty in “rolling over” its maturing debts or financing its deficit, and unemployment does not reach major proportions. In either case, the Federal Reserve has no choice but to administer to political dictate.

When the national credit and the national currency are “tapped” in order to maintain “full employment,” full employment might be maintained. The money market can be kept liquid indefinitely if the Treasury prints certificates and the Federal Reserve monetizes them. But what happens to the liquidity of the monetizer? The assets in the portfolio of the Federal Reserve System amountde facto to permanent investments. It makes little difference whether they consist of short- term certificates or of long-term bonds. In effect, they are as good as non-marketable consols. (The same holds for assets of the Federal Deposit Insurance Corporation, another fountain of pseudo-liquidity.) Disposing of as much as ten percent of the portfolio would “wreck” the credit markets. An over-indebted Treasury, one in deficit at that, cannot redeem the one kind or the other, but is bound to resort recurrently to more monetization.

By slow attrition, the result is likely to be the same as in the case of outright money-printing by the government itself. The old-fashioned technique of paper money inflation “worked” faster than its modern, seemingly less reprehensible counterpart that camouflages the production of fiat money by channelling it through the money market and the central bank. The latter’s liquidity consists exclusively of its gold reserve that tends to decline in proportion to its liabilities. The attrition of the gold reserve accelerates when the gathering of inflationary expectations induces non-resident owners of dollar balances to withdraw them (with residents joining, too). There can be little doubt of the final outcome, unless the process is brought to a halt.

SUGGESTIONS FOR FURTHER READINGS

K. Kock. A Study of Interest Rates. (London, King, 1929).

Analyzes the policies of leading central banks and of the Federal Reserve System, showing that and how illiquidity of commercial banks makes those policies ineffective.

A. A. Berle, Jr. and V. J. Pederson. Liquid Claims and National Wealth. (New York, MacMillan, 1934).

Interesting study of marketing methods making for “artificial” liquidity of claims. However, it confuses “liquidity” with wholesale liquidation, and liquidation of assets (securities) with that of currency (deposits).

Waldo Mitchell. Uses of Bank Credit. (University of Chicago Press, 1925).

Represents the “shiftability” point of view, denying the relevance of “liquidity” problems-which it overlooks.

C. F. Dunbar and 0. M. W. Sprague. The Theory and History of Banking. (New York, Putnams, 1917).

An introductory approach, especially Chapters VI and XII, to the problem of liquidity as involved in the maintenance of the gold standard, and to the causes of business cycles. Unfortunately, these most important issues connected with liquidity are hardly more than touched upon in recent Anglo-American literature.

Additional Suggestions:

Courtney C. Brown, Liquidity and Instability, New York, Columbia University Press, 1940

Antal E. Fekete, Borrowing Short and Lending Long: Illiquidity and Credit Collapse, C.M.R.E. Monograph Number 38, Greenwich, Connedicut, 1983

Caroline Whitney, Experiments in Credit Control, New York, 1934

More complete bibliographies can be found in the works listed above.

Seguros de salud, Mas-Colell y el argumento liberal-libertario (II)

“The mathematical prerequisites for use of the book are basic knowledge of calculus, some familiarity with linear algebra and a grasp of the elementary aspects of probability”

MAS-COLELL, ANDREU, MICHAEL DENNIS WHINSTON, AND JERRY R. GREEN. MICROECONOMIC THEORY

En una entrada anterior explicamos la mayor crítica de Mas-Colell a lo que él llma el argumento libertario a favor de un mercado de seguros sanitarios y lo hemos diferenciado de la falla de mercado que es la selección adversa. En esta entrada intentaremos responder a la crítica de Mas-Colell.


El punto de partida

El primero que podemos apreciar del modelo que presenta Mas-Colell es que las condificiones de eficiencia (seguro universal eficiente) depende del punto de partida del modelo. El modelo parte de una población homogénea, de ahí que todo el mundo se asegure a las mismas primas. Si además suponemos que tm = 0, no se recibe ningún subsidio cuando se está enfermo, a parte del tratamiento (en t = 2), la prima que pagarán todos los individuos en t = 1 ser´´a de αc (por la condición: tb + tm = -αc). Ahora, introduciremos una pequeña modificación al modelo que Mas-Colell presenta: una población heterogénea de partida. Sabemos que la población nunca es homogénea, sabemos que los individuos tienen diferencias apreciables entre sí de forma que un punto de partida más realista incorporará individuos con mayor o menor probabilidad de enfermar. Tenemos dos grupos de igual tamaño con diferentes probabilidades: 0 < α1 < 1 i 0 < α2 < 1. Si los diferentes grupos son suficientemente grandes para aplicar la ley de los grandes números, en t = 2 tendremos una fracción de la población  (α1 + α2)/2 enferma y las primas (con la condición anterior de tm = 0) para el primero grupo serán de α1c y para el segundo grupo α2c. Ahora imaginemos que en vez de dos grupos, tenemos tres gruòs de igual tamaño con las siguientes probabilidades de enfermar: 0 < α1 < 1, 0 < α2 < 1 i 0 < α3 < 1. Si los grupos son suficientemente grandes para aplicar la ley de los grandes números, en t = 2 tendremos una fracción de la población (α1 + α2 + α3)/3 enferma y las primas para el primer grupo serán de α1c, para el segundo grupo de α2c y para el tercer grupo de α3c. Como podemos ver, un mercado competitivo de seguros médicos, fija las primas a cada individuo según el coste esperado de los gastos médicos de este, que dependerán de su probabilidad de enfermar (ya que el coste es constante y como Mas-Colell asumimos una enfermedad y un nivel de tratamiento). Hará falta tener en cuento esto cuando analicemos el benchmark que utiliza Mas-Colell (el de la población homogénea).

La contratación ex ante: el salto de la homogeneidad a la heterogeneidad

En el post anterior hemos visto que Mas-Colell utiliza en su modelo de benchmark, el que utilizará para juzgar los resultados del argumento libertario. Como bien nota Mas-Colell:

La aplicación del principio libertario sobre el trasfondo de una población homogénea tendrà el resultado previsto, es decir, la cobertura universal eficiente, sota una condición: que el mercado (competitivo) y la contratación de póliza de seguros tenga lugar antes de la resolución de incertidumbre, es a decir, en un instante suficientemente ex ante para que todos los agentes econòmicos sean autènticamente simètricos (es decir, iguales). En efecto, si así es, entonces todos los ciudadanos adquieren en el mercado competitivo la misma cantidad de pólizas de seguros (simplemente por simetría: son todos iguales); en particular, todos ellos pagarán y recibirán lo mismo. Además, por el bien conocido teorema de la mano invisible (o, más técnicamente, por el <>) el resultado final, es a decir, el equilibrio competitivo del mercado, es eficiente. Sin embargo, en nuestro modelo simplificado, hemos visto que hay una sola asignación de recursos que es eficiente ex ante y por la cual las transferencias monetarias de todos los ciudadanos son idénticas: es la que resulta de la cobertura universal eficiente.

Andreu Mas-Colell

Como podemos ver el mercado libertario no tiene ningún problema… a menos que se modifique arbitrariamente el punto de partida, recordemos que: “La aplicación del principio libertario sobre el trasfondo de una población homogénea tendrà el resultado previsto, es decir, la cobertura universal eficiente, sota una condición: que el mercado (competitivo) y la contratación de póliza de seguros tenga lugar antes de la resolución de incertidumbre…”. Justamente Mas-Colell critica que un mecado competitivo, delante de una población de partida no homogénea es ineficiente ya que no cumple las condiciones de eficiencia que él encuentra partiendo de una población homogénea!

Rothschild i Stiglitz (1976) y la condición de eficiencia de que todo el mundo debe ser tratado.

Una vez hemos visto que Mas-Colell está aplicando la condición de eficiencia de poblaciones homogéneas para juzgar una población no homogénea, debemos ver que condiciones de eficiencia debería cumplir una población no homogénea En una entrada de blog, por temes prácticos y más por el trabajo que lleva, no haremos la demostración formal de las condiciones bajo las cuales el mercado es eficiente, aunque si tenemos mercados completos (información perfecta [en este caso la información exacta sobre la probabilidad de enfermar de cada agente y su condición médica en el momento de asegurarse], ausencia de externalidades, costes de transacción, etc), que todos los agentes toman los precios como dados y la no saciedad local, se cumple el primer teorema fundamental de la economía del bienestar, el mercado llega a un resultado Pareto óptimo (capítulo 19 “General Equilibrium Under Uncertainty” haciendo referencia al capítulo 16 “Equilibrium and it’s Basic Welfare Properties” de MICROECONOMIC THEORY) y implícitamente Mas-Colell asume estas premisas en su crítica. En el artículo de Rothschild y Stiglitz (1976) “Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Competition” (donde exponen la falla de mercado de la selección adversa, que no es lo que dice Mas-Colell, justamente a la entrada anterior vemos como el progreso tecnológico y médico soluciona este problema). En el artículo de Rothschild y Stiglitz, vemos que el equilibrio eficiente, el que se expone antes de entrar en un mercado con información imperfecta, es cuando las primas para cada individuo equivalen al coste esperado de enfermar, que se encuentra en el apartado 1.5 “Equilibrium with Identical Costumers” de donde también parte Mas-Colell, pero podemos observar en una nota de página:

5. The analysis is identical if individuals have different p‘s, [probabilitat de tenir un accident] but companies know the accident probabilities of their customers. The market splits into several submarkets-one for each different p represented. Each submarket has the equilibrium described here.

Rothschild & Stiglitz (1976)

Justamente en este apartado, donde todos los consumidores son iguales, no hay asimetria informativa para la aseguradora y en caso de que se conocieran, habría diferentes sub-mercados, uno para cada grupo de riesgo que comparte las mismas propiedades que Rothschild y Stiglitz (1976) especifican en su apartado 1.5. En este caso, que como hemos visto en la cita de la nota de página también puede englobar casos donde haya diferentes probabilidades pero estas sean conocidas por las aseguradoras:

In equilibrium each costumer buys complete insurance at actuarial odds.

Rothschild & Stiglitz (1976)

Como hemos podido ver en la primera parte de esta entrada, el mercado competitivo (con información perfecta) sigue el criterio que acabamos de citar, en el caso donde Rothschild & Stiglitz (1976) encuentran que no hay selección adversa. Si no añadimos otras condiciones que generen nuevas fallas de mercado, que Mas-Colell no lo hace, el hecho que algunas personas no puedan comprar el seguro, es decir, que esté fuera de su restricción presupuestaria, no implica ningún tipo de ineficiencia (de hecho implica una violación de lo que asume Mas-Colell, de que c es suficientemente baja para que todo el mundo sea tratado), pueden haber equilibrios Pareto óptimos que no son nada equitativos. Que en el mercado el precio de las primas sea equivalente al coste esperado de enfermar, es justamente (en ausencia de externalidades, etc) condición necesaria de eficiencia. Bajo las condiciones previamente especificadas, el mercado se comporta o sigue la misma regla que el modelo de Mas-Colell, las primas son el coste esperado de enfermar. Lo que comentamos aquí es que el mercado bajo una población heterogénea sigue la misma dinámica que expone Mas-Colell para el caso eficiente, lo único que Mas-Coell hace un salto de una población homogénea a una que no lo es para juzgar el mercado, cuando este parte de la segunda, imponiendo las condiciones de eficiencia que encuentra partiendo de la población homogénea.

 En resumen, encuentro que la crítica de Mas-Colell a lo que llama el argumento libertario no es exitosa ya que intenta juzgar los resultados de un mercado donde la población es heterogénea con los resultados que encuentra partiendo de una población homogénea.

Assegurança de salut, Mas-Colell i l’argument liberal-llibertari (II)

“The mathematical prerequisites for use of the book are basic knowledge of calculus, some familiarity with linear algebra and a grasp of the elementary aspects of probability”

MAS-COLELL, ANDREU, MICHAEL DENNIS WHINSTON, AND JERRY R. GREEN. MICROECONOMIC THEORY

En una entrada anterior vam explicar la major crítica de Mas-Colell al que ell anomena l’argument llibertari a favor d’un mercat d’assegurances sanitàries i l’hem diferenciat de la falla de mercat que es la selecció adversa. En aquesta entrada intentarem respondre a la crítica de-Mas Colell.


El punt de partida

El primer que podem apreciar del model que presenta Mas-Colell es que les seves condicions d’eficiència (assegurança universal eficient) depenen del punt de partida del model. El model parteix d’una població homogènia, d’aquí que tothom s’asseguri a les mateixes primes. Si a demés suposem que td = 0, no es rep cap subsidi quan s’està malalt (en t = 2) la prima que pagaran tots els individus en t = 1 será de αc (per la condició: tb + td = -αc). Ara, introduirem una modificació al model que Mas-Colell presenta: una població heterogènia de partida. Sabem que la població mai es homogènia, sabem que els individus tenen diferències apreciables entre sí de forma que un punt de partida més realista inclourà a individus amb major o menor probabilitat d’emmalaltir. Tenim dos grups d’igual tamany amb diferents probabilitats: 0 < α1 < 1 i 0 < α2 < 1. Si els diferents grups són suficientment grans per aplicar la llei dels grans nombres, en t = 2 tindrem una fracció de la població  (α1 + α2)/2 malalta i les primes (amb la condició anterior de td = 0) per al primer grup seràn de α1c i pel segon grup α2c. Ara imaginem que en lloc dos grups, tenim tres grups d’igual tamany amb les següents probabilitats de posar-se malalt:  0 < α1 < 1, 0 < α2 < 1 i 0 < α3 < 1. Si els grups són suficientment grans per aplicar la llei dels grans nombres, en t = 2 tindrem una fracció de la població (α1 + α2 + α3)/3 malalta i les primes per al primer grup serà de α1c, pel segon grup de α2c i per al tercer grup de α3c. Com podem veure, un mercat competitiu d’assegurances mèdiques carrega les primes a cada individu segons el cost esperat de les despeses mèdiques d’aquest, que dependran de la seva probabilitat d’emmalaltir (ja que el cost es ja que com Mas-Colell assumim una malaltia i un nivell de tractament). Caldrà tenir molt en compte això quan analitzen el benchmark que utilitza Mas-Colell (el de població homogènia).

La contractació ex ante: el salt de l’homogèneitat a la heterogeneitat

En el post anterior hem vist que Mas-Colell utilitza en el seu model de benchmark, el que utilitzarà per jutjar els resultats de l’argument llibertari. Com bé nota Mas-Colell:

“L’aplicació del principi lliberari sobre el trasfons d’una població homogénia tindrà el resultat previst, és a dir, la cobetura universal eficient, sota una condició: que el mercat (competitiu) i la contractació de la pòlissa d’assegurança tinguin lloc abans de la resolució de la incertesa, és a dir, en un instant prou ex ante perquè tots els agents econòmics siguin autènticament simètrics (es a dir, iguals). En efecte, si així és, aleshores tots els ciutadans adquiriran en el mercat competitiu la mateixa quantitat de pòlises d’assegurances (simplement per simetria: són tots iguals); en particular, tots ells pagaran i rebran el mateix. A més, pel ben conegut teorema de la mà invisible (o, més tècnicament, pel <<primer teorema fonamental de l’economia del benestar>>) el resultat final, és a dir, l’equilibri competitiu del mercat, és eficient.”

Andreu Mas-Colell

Com podem veure el mercat llibertari no té cap problema… a menys que es canviï arbitrariament el punt de partida, recordem que: “L’aplicació del principi lliberari sobre el trasfons d’una població homogénia tindrà el resultat previst, és a dir, la cobetura universal eficient, sota una condició: que el mercat (competitiu) i la contractació de la pòlissa d’assegurança tinguin lloc abans de la resolució de la incertesa…”. Justament Mas-Colell critica que en mercat competitiu, davant d’una població no homogènia es ineficient ja que no compleix les condicions d’eficiència si partim d’una població homogènia!

Rothschild i Stiglitz (1976) i la condició d’eficiència que tothom ha de ser tractat.

Una vegada ja hem vist que Mas-Colell està aplicant la condició d’eficiència de poblacions homogènies sobre una població no homogènia, hem de veure quina distribució eficient hauria de complir la població no homogènia. En un entrada d’un blog, per temes pràctics i més aviat per la feina que comporta, no farem una demostració formal de les condicions sota les quals el mercat es eficient, encara que si tenim mercats complets (informació perfecte [en aquest cas l’informació exacta sobre la probabilitat d’emmalaltir de cada agent i la seva condició mèdica en el moment d’assegurar-se], absència d’externalitats, costos de transacció, etc), que tots els agents prenen els preus com donats i la no sacietat local, es compleix el primer teorema fundamental del benestar, el mercat porta a un resultat Pareto òptim (capítol 19 “General Equilibrium Under Uncertainty” fent referència al capítol 16 “Equilibrium and it’s Basic Welfare Properties” de MICROECONOMIC THEORY) i implicitament Más-Colell assumeix aquestes premisses en la seva crítica. En l’article de Rothschild i Stiglitz, veiem que els equilibris eficients, el que primer exposen abans d’entrar el un mercat amb informació imperfecte, es on les primes per a cada individu equivalen al cost esperat de emmalaltir, que es en l’apartat 1.5 “Equilibrium with Identical Costumers” d’on Mas-Colell també parteix, però podem observar en una nota de pàgina:

5. The analysis is identical if individuals have different p‘s, [probabilitat de tenir un accident] but companies know the accident probabilities of their customers. The market splits into several submarkets-one for each different p represented. Each submarket has the equilibrium described here.

Rothschild & Stiglitz (1976)

En l’apartat on tots els consumidors són iguals, no hi ha asimetria informativa per a l’asseguradora i en cas de que es conseguéssin, hi hauria diferents sub-mercats, un per a cada grup de risc que comparteixen les mateixes propietats que Rothschild i Stiglitz (1976) especifiquen en el seu apartat 1.5. En aquest cas, que com hem vist en la cita de la nota de pàgina també pot englobar casos on hi hagi diferents probabilitats d’emmalaltir però aquestes són conegudes per part de les asseguradores:

In equilibrium each costumer buys complete insurance at actuarial odds.

Rothschild & Stiglitz (1976)

Com hem pogut veure en la primera part d’aquesta entrada, el mercat competitiu segueix el criteri que acabem de citar anteriorment, en el cas on Rothschild & Stiglitz (1976) troben que no hi ha selecció adversa. Si no afegim altres condicions que generin noves falles de mercat, que Mas-Colell no ho fà, el fet que algunes persones no es puguin comprar l’assegurança, es a dir, aquesta estigui fora de la seva restricció pressupostària, no implica cap tipus d’ineficiència (de fet implica una violació del que assumeix Más-Colell, que c es sempre suficientment baix perquè tothom sigui tractat), de fet hi poden haver equilibris Pareto òptims que no son gens equitatius. Sota les condicions prèviament especificades, el mercat es comporta igual que el model de Mas-Colell, les primes són el cost esperat d’emmalaltir, el que comentem aquí es que el mercat sota una població heterogènia segueix la mateixa dinàmica que el cas que exposa Mas-Colell per al cas eficient, l’únic que Mas-Colell fa un salt de una població homogènia a una que no ho és. Seguint Rothschild i Stiglitz (1976) el mateix criteri que Más Colell, ja noten que simplement, quan hi ha diferents probabilitats, el mercat es divideix en diferents sub-mercats per a cada grup de risc, com hem exposat que faria el mercat a la primera part.

 En resum, trobo que la crítica al que Mas-Colell anomena l’argument llibertari fa aigues ja que intenta jutjar els resultats d’un mercat on la població és heterogènia amb els criteris d’eficiència que “troba” partint d’una població homogènia.

Seguros de salud, Mas-Colell y el argumento liberal-llibertario (I)

“The mathematical prerequisites for use of the book are basic knowledge of calculus, some familiarity with linear algebra and a grasp of the elementary aspects of probability”

MAS-COLELL, ANDREU, MICHAEL DENNIS WHINSTON, AND JERRY R. GREEN. MICROECONOMIC THEORY

En 1994 se nos regaló la creación de un libro ‘ANÀLISI ECONÒMICA DE LA SANITAT’ patrocinado por la administración sanitaria de Cataluña, que es una recolección de escritos por una serie de autores, entre ellos Andreu Mas-Colell, Xavier Freixes y Joseph Stiglitz. Todos los escritos se desarrollan alrededor de la economía de la salud, aunque hay otros que vierten de forma más general sobre la cuestión del Estado del Bienestar.

Mas-Colell (PhD, University of Minnesota, 1974), es actualmente consejero de Economía y Conocimiento de la Generalitat de Cataluña, ha sido un investigador muy prolífico, la mayoría de sus contribuciones giran alrededor del equilibrio neowalrasiano. Mas-Colell ha sido professor a Harvard, Berkley y ahora lo es de la Universidad Pompeu Fabra (en excedencia). Fue editor de Econometrica entre 1988 y 1992, una de las revistas más importantes de economía a nivel mundial y fundador de la Barcelona Graduate School of Economics. Es fellow de la Econometric society y fue su presidente en 1993. La publicación del libro mencionado, justamente fe un año antes que Mas-Colell publicara, junto con otros dos autores, su magnum opus, el libro que se convertiría en el manual de microeconomía de referencia mundial: MICROECONOMIC THEORY, del cual la cita inicial de esta entrada es un chiste de mal gusto. Recientemente, la University of Chicago le ha otorgado un doctorado honoris causa.

El argumento libertario y la critica de Mas-Colell

El escrito de Mas-Colell en la obra mencionada tiene por título: ‘Sobre el carácter obligatorio y universal del seguro de salud’ [versión en catalan y una traducción no profesional al castellano], versa en especial sobre lo que él llama el argumento libertario (de libertarian en los EE.UU), de como no es exitoso y que remedios hay, casi siempre, des de un criterio de eficiencia.

Según define Mas-Colell, el argumento libertario consiste en:

Reducida a lo esencial, la posición libertaria mantendrá que una póliza de seguro médico sería como cualquier otra mercancía comerciable en mercados competitivos. Los agentes económicos son libres y soberanos para adquirirla. Si así lo hacen, quedan cubiertos en caso de crisis de salud, y al contrario si no lo hacen. Pero, para el libertario, esto no debería ser motivo de preocupación social. La enfermedad es un acto de Dios sobre el cual no se le puede hacer nada a priori y los gastos médicos son una elección consciente y meditada de un sujeto económico que ya sabía que tendría que asumir las consecuencias de la no adquisición de seguro médico en la eventualidad de enfermedad.

Mas-Colell cree que el argumento libertario es especialmente problemático por la dificultad de contratar un seguro ex ante (previamente a la resolución de incertidumbre).

Mas-Colell, empieza exponiendo un modelo simplificado de un mercado competitivo de seguros médicos y ver cuales son las condiciones de eficiencia:

Como el teórico es muy libre de hacer sus hipótesis, supondré, en consecuencia, que nos concierne una colectividad formada por individuos de idéntico nivel de riqueza y de renta. Para no introducir diferencias de forma indirecta, supondré además que los miembros de la sociedad son idénticos ex ante, es decir, antes de la resolución de incertidumbre. Ex post, es decir, después de la resolución de incertidumbre, los diferentes individuos, pueden, sin embargo, exhibir diferentes grados de salud.

Un modelo formal de simplicidad extrema podría ser el siguiente. Tenemos dos periodos, t = 1 y t = 2. Hay dos estados personales de salud: buena (b) y mala (m). El estado de salud se conoce al principio del segundo periodo. La probabilidad de ponerse enfermo es 0 < α < 1, y es independiente entre los diferentes individuos. Suponiendo que la población es suficientemente numerosa para la aplicación de la ley de los grandes números, sigue que al principio del periodo 2 una fracción α de la población necesitará atención médica. Para simplificar, supondré que hay un tratamiento médico disponible solo a dos niveles: o este se aplica o no se aplica. El coste de aplicación es c > 0 y, en principio, los enfermos pueden recibirlo o no. A fin de concentrarme en el problema que ahora nos interesa, supondré que además, que el estado de salud es suficientemente penoso, o que c > 0 es suficientemente bajo para que cualquier asignación eficiente de recursos implique que todos los enfermos reciben el tratamiento .

Bajo las hipótesis anteriores hay una sola asignación social de los recursos que sea eficiente y preserve la igualdad entre los agentes económicos (es decir, que no implique redistribuciones de riqueza). La podemos describir de la manera siguiente. En primer lugar todos los enfermos serán atendidos, a uno coste total de αc. En segundo lugar, a cada agente se le asignará el mismo esquema de pago/subsidios (tb,tm) donde la cantidad tb es interpretada como la cantidad que se tiene que pagar (si tb < 0) o recibir (si tb > 0) si el estado de salud es bueno y, similarmente, la cantidad tm es interpretada como la cantidad que se tiene que pagar (si tm < 0) o recibir (si tm > 0) si el estado de salud es malo (y se recibe tratamiento médico). Estas cantidades tienen que cumplir que tb + tm = -αc, a fin de financiar el sistema sanitario. Sujeto a esa condición, los valores de tb, tm  se determinarán por las condiciones habituales de eficiencia (es decir, igualando las utilidades marginales de la renta en los dos estados). Bajo las condiciones normales del problema siempre tendremos tb < 0 (es decir, en caso de buena salud se paga). En cuanto a tm es posible que tm < 0 (también se paga cuando se está enfermo) pero también que tm > 0 (es decir, además del tratamiento médico se reciba un subsidio). Este esquema de pagos/subsidios tiene la interpretación natural de un seguro de salud: todos los ciudadanos, que son idénticos, pagan una prima -tb con independencia de su estado de salud (es decir, pagada en la práctica antes que se conozca el estado de salud) y, en caso de enfermedad reciben tratamiento médico y la cantidad tb-tm (típicamente una cantidad no negativa). A este arreglo, que de ahora en adelante centrará nuestra atención, lo llamaré la cobertura universal eficiente..

El problema para el libertario consisten en la creciente dificultad de contratar ex ante y si el mercado intenta actuar bajo los criterios (de eficiencia) expuestos por Mas-Colell en la cita anterior para evitar la crítica, Mas-Colell comenta que habrá problemas de selección adversa. Aunque el problema de la selección adversa es de grran interés, avaluaremos la crítica inicial de Mas-Colell a la posición libertaria (en una futura entrada) en sus propios méritos, especialmente si el problema de la selección adversa solo se da cuando el mercado intenta evitar la crítica de Mas-Colell operando bajo las condiciones previamente especificadas.

Un seguro de salud, como todo seguro, debe su virtualidad a la posibilidad de ser adquirida antes de que en el lenguaje técnico llamaremos la resolución de incertidumbre. Por ejemplo, es evidente, tanto para el teórico libertario como para nosotros, que cuando se ha desarrollado una condición de hipertensión ya no es el momento de pretender adquirir un seguro de salud a primas razonables. Es decir -repitiendo el argumento de más arriba- con una población homogénea, la eficiencia económica requiere que el momento de decisión para el acto asegurador se dé cuando los agentes económicos aún son iguales ex ante y la separación entre favorecidos por la fortuna aún no haya empezado.

Pero es aquí precisamente donde hay el problema y la dificultad, para mí decisiva, del planteamiento libertario. Es dudoso que esta posibilidad de un momento original e indiferenciable haya existido nunca, pero el que es seguro es que el progreso de la tecnología y la ciencia medica hace que, cada vez más, este hipotético momento original de igualdad ex ante se retraiga al tiempo que precede el nacimiento. Nos acercamos al día en que seremos un libro abierto cuando nacemos, el día en que, por ejemplo, el análisis de una célula al nacer, o en estado fetal, proporcionará una radiografía mucho más precisa de la categoría de riesgo a la que pertenecemos. O mejor decir, en que tal examinación tendrá la posibilidad de catalogarnos en categorías de riesgo muy diferentes desde el instante mismo, o antes, de llegar al mundo. A mi me parece que este avance del conocimiento científico y de la tecnología es imparable y que pretender ponerle barreras sería una tarea de suma futilidad.

Cabe subrayar ese punto porque tiene una importancia fundamental: en las condiciones descritas más arriba la provisión vía mercados de servicios sanitarios no podía conseguir la cobertura universal eficiente (en particular, en el nuestro caso de una población homogénea) porque el hecho que los individuos en el momento de asegurarse ya son diferentes. En esta circunstancia quien pertenezca a las categorías desfavorables terminará pagando más y el resultado final no será, por lo tanto, la cobertura universal eficiente (podrá suceder que, aún así, los enfermos reciban tratamiento médico simplemente porque ellos mismos se lo pagan ex post, pero es evidente que este sistema de financiamiento se aleja mucho de la eficiencia ex ante, que es la que cuenta).

Para Mas-Colell, la solución es un seguro público obligatorio, que podría imponer la cobertura universal eficiente y a la vez hacer uso de la información adicional que revelan los avances tecnológicos y médicos.

La crítica de MAS-COLELL VS la selección adversa, no nos confundamos.

El argumento de Mas-Colell tiene a ver con la creciente dificultad de contratación ex ante, las aseguradoras cobrarán primas diferentes del modelo con la población homogénea (que es como aparenta una población si no se conocen las diferencias de riesgo entre individuos), la aseguradora tiene “demasiada” información y al cobrar primas diferentes no tiene porqué dar-se la cobertura universal eficiente: “En resumen: la disponibilidad, todo y en principio, de la información destruye la posibilidad de la cobertura universal eficiente”. El problema de la selección adversa es justo al contrario, para que esta se dé, es necesaria la existencia de información asimétrica, el comprador del seguro tiene que saber más sobre su salud que la aseguradora. Aún así, nos encontramos con esto: “He de advertir que el problema de información que estoy discutiendo es bien conocido en la literatura bajo el nombre del problema de la selección adversa, y la tesis que estoy manteniendo, es al fin y al cabo, que el progreso técnico está  agravando aceleradamentee este problema y lo está convirtiendo en un problema (por no decir el problema) de relevancia central en el análisis económico de la salud”.

STIGLITZ I ROTHSCHILD (1976)

STIGLITZ I ROTHSCHILD (1976)

Supongamos, de buena fe (sin acusar a Mas-Colell de error), que se refiere a si las compañías de seguros intentaran solucionar el problema que el menciona inicialmente, con precios sin discriminar por riesgo (como si no conocieran las diferencias de riesgo entre individuos, ignoran esta información) entonces se da selección adversa. No es el caso que el progreso tecnológico acentuara el problema directamente, al contrario, debería desaparecer ya que cada vez será más difícil que uno tenga información privada sobre su propia salud (de hecho el progreso tecnológico y medico del que habla Mas-Colell, parece obra de la función empresarial a la Kizner que soluciona la selección adversa, una falla de mercado):

En efecto, si las pólizas están disponibles a las primas justificadas ex ante, es a decir, a las que corresponden a las probabilidades conocidas (de hecho o por disciplina) por el asegurador, entonces las primas van a ser demasiado caras para los ciudadanos que saben que tienen estados de salud favorables y muy atractivas para los que saben que tiene estados de salud desfavorables. Los primeros se abstendrán, por tanto, de participar en el mercado. La cosa va más allá. Que la información sea en principio obtenible hace que, en el fondo, no importa si los ciudadanos saben o no saben, de hecho, su condición de salud. El asegurador deberá de contar con el hecho que, con los precios justificados ex ante, se producirá la abstención que acabamos de describir. Por tanto, hará falta que anticipe que la composición de sus clientes es desfavorable y, en consecuencia, deberá aumentar la prima. No desarrollaremos el bien conocido análisis de esta cuestión (ver Rothchild y Stiglitz 1976)…

En un futuro próximo, en una siguiente entrada, analizaremos críticamente este argumento de Mas-Colell.

Assegurança de salut, Mas-Colell i l’argument liberal-llibertari (I)

“The mathematical prerequisites for use of the book are basic knowledge of calculus, some familiarity with linear algebra and a grasp of the elementary aspects of probability”

MAS-COLELL, ANDREU, MICHAEL DENNIS WHINSTON, AND JERRY R. GREEN. MICROECONOMIC THEORY

El 1994, se’ns va regalar la creació d’un llibre ‘ANÀLISI ECONÒMICA DE LA SANITAT’ patrocinat per l’administració sanitària de Catalunya, que és una recol·lecció d’escrits per una sèrie d’autors, entre ells Andreu Mas-Colell, Xavier Freixes i Joseph Stiglitz. Tots els escrits es desenvolupen al voltant de l’economia de la salut, encara que n’hi ha d’altres que toquen de forma més general la qüestió de l’Estat del Benestar.

Mas-Colell (PhD, University of Minnesota, 1974), es actualment conseller d’Economia i Coneixement de la Generalitat de Catalunya, ha estat un investigador molt prolífic, la majoria de les seves contribucions han girat entorn l’equilibri neowalrasiá. Mas-Colell ha estat professor a Harvard, Berkley i ara ho és de la Universitat Pompeu Fabra (en excedència), va ser editor de Econometrica entre 1988 i 1992, una de les revistes més importats d’economia en l’àmbit mundial i fundador de la Barcelona Graduate School of Economics. És fellow de la Econometric Society i en va ser el seu president el 1993. La publicació del llibre mencionat, justament va ser un any abans que Mas-Colell publiqués, juntament amb dos altres autors, el seu magnum opus, el llibre que es convertiria en el manual de microeconomia de referència mundial: MICROECONOMIC THEORY, del qual la cita inicial d’aquesta entrada és un acudit de mal gust. Recentment se li ha otorgat un doctorat honoris causa per la University of Chicago.

L’argument llibertari i la crítica de MAS COLELL

L’escrit de Mas-Collel en l’obra mencionada a l’inici té per títol ‘Sobre el caràcter obligatori i universal de l’assegurança de salut’ [versió en català i una traducció no professional al castellà], versa en especial sobre el que anomena l’argument llibertari (de libertarian als EE.UU), de com no es exitós i quins remeis hi ha, gairebé sempre, des d’un criteri d’eficiència.

Segons defineix Mas-Colell, l’argument llibertari consisteix en:

Reduïda a l’essencial, la posició llibertària mantindria que una pòlissa d’assegurança mèdica seria com qualsevol altra mercaderia comerciable en mercats competitius. Els agents econòmics són lliures i sobirans per adquirir-la. Si ho fan així, queden coberts en cas de crisi de salut, i al contrari si no ho fan. És clar que qui no l’hagi adquirida patirà doblement en cas de malaltia: per la malaltia mateixa i pels més grans dificultats i cost de proveir-se d’atenció mèdica. Però, per al llibertari, això no hauria de ser motiu de preocupació social. La malaltia és un acte de Déu sobre el qual res no es pot fer a priori i les despeses mèdiques són una elecció conscient i meditada d’un subjecte econòmic que ja sabia que hauria d’assumir les conseqüències de la no adquisició d’assegurança en l’eventualitat de malaltia.

Mas-Colell creu que l’argument llibertari és especialment problemàtic per la dificultat de contractar una assegurança ex ante (previ a la resolució d’incertesa).

Mas-Colell, comença exposant un model molt simplificat d’un mercat competitiu d’assegurances mèdiques i veure quines són les condicions d’eficiència:

Com que el teòric és molt lliure de fer les seves hipòtesis, suposaré, en conseqüència, que ens concerneix una col·lectivitat formada per individus d’idèntic nivell de riquesa i de renda. Per no introduir diferències de forma indirecta, suposaré, a més, que els membres de la societat són idèntics ex ante, és a dir, abans de la resolució de la incertesa, els diferents individus, poden tanmateix exhibir diferents graus de salut.

Un model formal de simplicitat extrema podria ser el següent. Tenim dos períodes, t = 1 i t = 2. Hi ha dos estats personals de salut: bona (b) i dolenta (d). L’estat de salut es coneix al començament del segon període. La probabilitat de posar-se malalt és 0 < α < 1, i és independent entre els diferents individus. Suposant que la població es prou nombrosa per a l’aplicació de la llei dels grans nombres, se segueix que al principi del període 2 una fracció α de la població necessitarà atencions mèdiques. Per simplificar, suposaré que hi ha un tractament mèdic disponible només a dos nivells: o aquest s’aplica o no s’aplica. El cost d’aplicació és c > 0 i, en principi, els malalts poden rebre’l o no. A fi de concentrar-me en el problema que ara ens interessa, suposaré, a més, que l’estat de la malaltia és prou penós, o que c > 0 és prou baix, perquè qualsevol assignació eficient de recursos impliqui que tots els malalts reben tractament.

Sota les hipòtesis anteriors hi ha una sola assignació social de recursos que sigui eficient i que preservi la igualtat entre els agents econòmics (és a dir, que no comporti redistribucions de riquesa). La podem descriure de la manera següent. En primer lloc tots els malalts seran atesos, a un cost total de αc. En segon lloc, a cada agent se li assignarà el mateix esquema de pagaments/subsidis (tb,td) on la quantitat tb és interpretada com la quantitat que s’ha de pagar (si tb < 0) o rebre (si tb > 0) si l’estat de salut és bo i, semblantment, la quantitat td és interpretada com la quantitat que s’ha de pagar (si td < 0) o rebre (si td > 0) si l’estat de salut es dolent (i es rep tractament mèdic). Aquestes quantitats han de complir tb + td = -αc, a fi de finançar el sistema sanitari. Subjecte en aquesta condició, els valors de tb, td  es determinaran per les condicions habituals d’eficiència (és a dir, igualant les utilitats marginals de la renda en els dos estats). Sota les condicions normals del problema sempre tindrem tb < 0 (és a dir, en cas de bona salut es paga). Pel que fa td és possible que td < 0 (també es paga quan s’està malalt) però també que td > 0 (és a dir, a més del tractament mèdic es rep un subsidi). Aquest esquema de pagaments/subsidis té la interpretació natural d’una assegurança de salut: tots els ciutadans, que són idèntics, paguen una prima -tb amb independència de l’estat de salut (és a dir, pagada en la pràctica abans que es conegui l’estat de salut) i, en cas de malaltia reben tractament i la quantitat tb-td (típicament una quantitat no negativa). A aquest arranjament, que d’ara endavant centrarà la nostra atenció, l’anomenaré la cobertura universal eficient.

El problema per al llibertari consisteix en la creixent dificultat contractació ex ante i si el mercat intenta actuar sota els criteris exposats de Mas Colell en la cita anterior per evitar la crítica, Mas-Colell comenta que hi haurà problemes de selecció adversa. Tot i que el problema de la selecció adversa és de gran interès, avaluarem la crítica inicial del Mas Colell a la posició llibertaria (en el següent post) en els seus propis mèrits, especialment si el problema de la selecció adversa es dóna només quan el mercat intenta evitar la crítica de Mas-Colell operant sota les condicions prèviament especificades.

Una assegurança de malaltia, com d’altra banda tota assegurança, deu la seva virtualitat a la possibilitat de ser adquirida abans del que en llenguatge tècnic anomenarem la resolució d’incertesa. Per exemple, és evident, tant per al teòric llibertari con per a nosaltres, que quan s’ha desenvolupat una condició d’hipertensió ja no és el moment de pretendre adquirir una assegurança de salut a primes raonables. Es a dir -repetint l’argument de més amunt- amb una població homogènia, l’eficiència econòmica requereix que el moment de decisió per l’acte assegurador es doni quan els agents econòmics encara són iguals i ex ante i la separació entre els afavorits per la fortuna encara no hagi començat.

Però és aquí precisament on hi ha el problema i la dificultat, per a mi decisiva, del plantejament llibertari. Es dubtós que aquesta possibilitat d’un moment original i diferenciable hagi existit mai, però el que és segur és que el progrés de la tecnologia i la ciència mèdica fa que, cada vegada més, aquest hipotètic moment original d’igualar ex ante es retrotregui als temps que precedeixen el naixement. Ens acostem al dia en què serem un llibre obert quan naixem, el dia en què, per exemple, l’anàlisi d’una cèl·lula al néixer, o en estat fetal, proporcionarà una radiografia molt precisa de la categoria dels riscs a la qual pertanyem. O millor dir, en què tal examen tindrà la possibilitat de catalogar-nos en categories de risc molt diferents des de l’instant mateix, o abans, de venir al món. A mi em sembla que aquest avença de coneixement científic i de la tecnologia és imparable i que pretendre posar-li barreres seria una tasca d’una futilitat summa.

Cal subratllar aquest punt perquè té una importància fonamental: en les condicions descrites més amunt la provisió via mercats de serveis de salut no pot aconseguir la cobertura universal eficient (en particular, en el nostre cas de la població homogènia) perquè de fet els individus en el moment d’assegurar-se ja són diferents. En aquesta circumstància qui pertanyi a categories desfavorables de risc acabarà pagant més i el resultat final no serà, per tant, la cobertura universal eficient (podrà succeir que, tot i això, els malalts rebin tractament mèdic perquè ells mateixos s’ho paguen  ex post, però és evident que aquest sistema de finançament s’allunya molt de l’eficiència, ex ante, que és la que compta).

Per a Mas-Colell, la solució és una assegurança pública obligatòria, que podria imposar la cobertura universal eficient i a l’hora fer ús de la informació addicional que revelen els avenços tecnològics i mèdics.

La crítica de MAS-COLELL VS la selecció adversa, no confonem.

L’argument de Mas-Colell té a veure que amb la creixent dificultat de contractació ex ante, les asseguradores cobraran primes diferents del model amb la població homogènia (que és com sembla una població si no es coneixen les diferències de riscos entre individus), l’asseguradora té “massa” informació i al cobrar diferents primes no té perquè donar-se la cobertura universal eficient: “En resum: la disponibilitat, tot i en principi, de la informació destrueix la possibilitat de la cobertura universal eficient”. El problema de la selecció adversa és just al contrari, perquè aquesta es doni es necessari que hi hagi informació asimètrica, el comprador d’assegurança ha de saber més sobre la seva salut que l’asseguradora. Tot i així ens trobem això: “He d’advertir que el problema d’informació que estic discutint és ben conegut en la literatura sota el nom del problema de la selecció adversa, i la tesi que estic mantenint és, al capdavall, que el progrés tècnic està agreujant acceleradament aquest problema i l’està convertint en un problema (per no dir, el problema) de rellevància central en l’anàlisi econòmica de la salut.

STIGLITZ I ROTHSCHILD (1976)

STIGLITZ I ROTHSCHILD (1976)

Suposarem, de bona fe (sense acusar a Mas-Colell d’equivocació), que es refereix a si les companyies d’assegurances intenten solucionar el problema que ell menciona inicialment, posant preus sense discriminar per risc (com si en no coneguessin les diferències de risc entre individus, ignoren aquesta informació) llavors es dóna selecció adversa. No és el cas que el progrés tecnològic accentuï el problema directament, al contrari, hauria de desaparèixer ja que cada vegada serà més difícil que un tingui informació privada sobre la seva pròpia salut (de fet el progrés tecnològic i mèdic del qual parla Mas-Colell, sembla obra de la funció empresarial de Kizner que soluciona la selecció adversa, una falla de mercat):

En efecte, si les pòlisses estan disponibles a les primes justificades ex ante, és a dir, a les que corresponen a les probabilitats conegudes (de fet o per disciplina) per l’assegurador, aleshores les primes seran massa cares per als ciutadans que saben que tenen estats de salut favorables i molt atractives per als qui saben que tenen estats de salut desfavorables. Els primers s’abstindran, per tant, de participar en el mercat. Las cosa va més enllà. Que la informació sigui en principi obtenible fa que, en el fons no importi si els ciutadans saben o no saben, de fet, la seva condició de salut. L’assegurador haurà de comptar amb el fet que, envers els preus justificats ex ante, es produirà l’abstenció que acabem de descriure. Per tant, caldrà que anticipi que la composició dels seus clients és més desfavorable i, en consqüència, haurà d’augmenar la prima. No desenvoluparem la ben coneguda conseqüència d’aquesta questió (vegeu Rothchild i Stiglitz 1976)…

En un futur pròxim, en una següent entrada analitzarem críticament aquest argument de Mas-Colell.