The impact of illiquidity risk for the Nordic markets

ABSTRACT

In this study, we propose a new measure of illiquidity for small, open stock markets; dollar zero-returns. Compared with other commonly used measures of illiquidity, the dollar zero-return produces the highest anomalous return across all four Nordic markets. In testing the pricing implication of the proxies of illiquidity, we use liquidity adjusted asset pricing models suggested in the literature for a panel of 25 size-related portfolios for the Nordic markets. Our results show that the only illiquidity measure that gives a significant positive effect across all Nordic markets is the dollar zero-return. Our results also show that the illiquidity mimicking portfolio factor, constructed through dollar zero-return, is the only factor showing a significant premium across different specifications, and its pricing remain significant also when the effect of the level of illiquidity (constructed through all measure of illiquidity) and of size is netted out.

RESUMEN

En este estudio se propone un nuevo indicador de iliquidez para mercados de valores pequeños y abiertos; rendimiento cero en dólares. Comparado con otros indicadores de iliquidez comúnmente usados, el rendimiento cero en dólares produce el rendimiento anómalo más alto en los cuatro mercados nórdicos. Con el propósito de verificar los efectos en valoración de las variables aproximadas de iliquidez, se utiliza el modelo de valoración de activos financieros ajustado por liquidez (LCAPM), propuesto por Acharya and Pedersen (2005) y se incluye el modelo de Liu (2006) para un panel de datos de veinticinco portafolios de tamaño similar en los mercados nórdicos. Los resultados muestran que el único indicador de iliquidez que presenta un efecto positivo significativo en todos los mercados nórdicos es el rendimiento cero en dólares. Los resultados también muestran que el factor de portafolio que imita la iliquidez, construido con el rendimiento cero en dólares, es el único que muestra una prima significativa para diferentes especificaciones, y su valoración se mantiene significativa cuando se descuenta el efecto del nivel de iliquidez (construido para todos los indicadores de iliquidez) y de tamaño.

KEYWORDS:

• Market-wide illiquidity
• return spread
• market model

JEL CLASSIFICATION:

• G11
• G12
• G15

PALABRAS CLAVES:

• iliquidez para todo el mercado
• margen sobre el rendimiento
• modelo del mercado

Acknowledgments

This paper has benefitted from comments and suggestions from participants at Hanken School of Economics departmental seminars in 2012 and 2013. The discussions with Prof. Johan Knif and Dr. Nader Virk are greatly acknowledged.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Absolute market returns as profitability of trade can be through buying and selling, as investor benefits neither from buying nor from selling if absolute change in market return is small.

2. Bekaert et al. (2007) used this monthly zero-return measure to examine the pricing implications of illiquidity risk. In addition, Goyenko et al. (2009) showed that this zero-return measure is related to a finer measure of illiquidity when estimated using high frequency data.

3. The data is taken from IMF website on 9/22/2017; the data source is Coordinated Direct Investment Survey CDIS. The provided estimates are the annual average of ratio of non-resident investment in dollar over total market capitalisation of the respective markets for the time series of 2009–2015.

4. Even for local investors the dollar zero-return is better measure of illiquidity, as it gives higher value of illiquidity to the consecutively non-trading stocks. As the consecutive non-trading reduces a chance of getting higher returns through random exchange rate fluctuations.

5. There is another additional benefit of measuring illiquidity is dollar zero-returns, as explained in Bekaert et al. (2007) that if the number of days of zero returns (non-trading) are the same for two stocks, then the illiquidity is more pronounced for the stock with consecutive non-trading days.

6. We define a no change to be when the prices in dollars of the stock are same till two decimals for two consecutive days. Whereas, prices in dollars for two consecutive days are same for that stock which has zero-return in local market and this event is accompanied with almost no change in $/local exchange rate.

7. This independence means that zero return in dollar denomination is a product of probability of zero return in local equity market and of no currency return. This observation allocates higher illiquidity to those stocks which have continuous long period of non-trading. To illustrate this point, we take a hypothetical example of two stocks that are not traded for 10 days, wherein the first case of non-trading days is randomly distributed and in the second case non-trading days are consecutive. For an ease of exposition we suppose that the probability of zero returns in the equity market for any given day and of the incidence of no change in currency return is 1/2 .As these events are independent, as per Figure 1, then the probability of zero returns in both markets is 1/4 and the simple expectation of 10 zero return days in both markets for the first stock will be 10 × 1/4 = 2.5. For second stock, on the other hand, the expectation of 10 zero return days in both markets will be 10 × 1/2 = 5, because in the equity market these zero return days occur consecutively and we know that already with probability 1, therefore zero returns on both markets is just the expectation of no change is local/$ exchange rates. As these expected numbers of zero return days are found in the numerator of Equation (2), the second stock, with a longer non-trading period, is the more illiquid of the two stocks. So, to estimate the Equation (2) we download the day end prices in dollar denomination for each Nordic market till two decimal points and no change in these prices for two days mean zero dollar return for that day.

8. For interested readers we refer to the paper of Fong et al. (2017) for a detailed analysis of the FTH measure in comparison to other proxy measure of liquidity.

9. Lesmond et al. (1999), Hasbrouck (2009), Corwan and Schultz (2012) propose new measures of illiquidity and show the effectiveness of their constructed measure by comparing it with a proxy measure of the effective spread proposed by Roll (1984).

10. In the literature, many authors have converted a positive covariance into a negative one (see, for example, Harris (1990) and Lesmond et al. (1999)). Doing the same in our analysis would produce some counterintuitive results.

11. Size is defined as the number of shares outstanding multiplied by the day end prices.

12. High correlation between market and illiquidity-related betas are reported in Acharya and Pedersen (2005) paper.

13. Except for Finland, where ZL$ correlation with Roll is −0.0474.

14. We also use an alternative specification of the Amihud measure where we estimate the illiquidity only for the stocks that has been traded at least 15 days in a given month.

15. We exclude the Roll-Spread measure from this analysis due to a lot of non-measurable observations.

16. We have not conducted any comparison between level and risk associated with illiquidity effect. Our main motivation in this paper is to test the illiquidity effect by using a newly proposed measure of illiquidity.

17. As we have seen in Table 2, that except for ZR$ there is no other measure of illiquidity which is consistent across different Nordic markets to establish higher premium linked with illiquidity effect.