Words Matter: Effects of Semantic Similarity of Monetary Policy Committee’s Decision on Financial Market Volatility
The objective of the paper is to study the effects of semantic similarity of the Bank of Thailand’s press releases on volatility of financial markets in Thailand from 2010-2018. The Natural Language Processing (NLP) is employed to construct the semantic similarity from 72 press releases. The semantic similarity represents the public signal that the central bank delivers to the public in the framework of a Keynesian beauty contest game.
The semantic similarity of MPC press releases significantly reduce the volatility in 1-month, 3-month, 10-year and 15-year government debt securities. Findings imply that relatively similar language in the MPC press releases reduces the volatility in short-term and long-term bonds. Effects of semantic similarity matter most in the volatility of 10-year bond yield. However, effects of semantic similarity are insignificant in both equity and foreign exchange markets.
Gaining from Digital Disruption: the Thai Financial Landscape in the Digital Era
This paper examines competitiveness of the Thai financial sector through the dimensions of depth, access, efficiency, and stability, as compared to peers. The paper finds that while the Thai financial sector compares reasonably well with peers in most dimensions, it does not fare well in term of SME access to bank credit. Using Panzar-Rosse H-Statistic, the paper also examines competition in the Thai banking sector and finds that the level of competition in the Thai banking sector is consistently high over the sample period. The results raise the question: Why does SME access to bank credit remain low, despite high level of competition in the banking sector? This puzzle is important since SMEs are a key driver of the Thai economy. Reviewing results from various studies and interviews with SMEs and bank credit officers, the paper identifies several bottlenecks in the SME lending process that may lead to market failures. Using data from 1.29 million individual SME loan contracts obtained from 15 Thai commercial banks, and six Specialized Financial Institutions (SFIs), the paper finds that only a few banks attempt to penetrate SMEs at the lower tiers of loan size and income. Although SME lending by SFIs are found to be a good complement to SME lending by banks, the fact remains that fewer than half of SMEs in Thailand have loans from these financial institutions. The paper then discusses how several initiatives have been attempted to harness the power of technology and data to help improve SME access to finance, whether from traditional banks or other types of players. Lessons from the case of SME financing and from other segments of financial sectors in selected countries are then drawn into common themes that might help guide the design of financial landscape in the digital era.
Mapping Thailand’s Financial Landscape: A Perspective through Balance Sheet Linkages and Contagion
This paper conducts in-depth profiling of players and interlinkages in the Thai financial system based on sectoral balance sheet data and disaggregated supervisory data on banks and mutual funds. Several aspects of Thailand’s financial landscape have been documented. We find that financial interconnectedness has risen and become more complex, with the financial landscape increasingly tilted toward non-bank intermediaries. Network topology suggests a segmented landscape, with the presence of a core cluster where key players including households, firms, large domestic banks, and mutual funds of large banks’ asset management arms are located, indicating their tight interconnections. Leveraging on entity-level balance sheet profiles, we develop a stress-testing framework that is based on a network model of financial contagion. Two types of shocks are studied. For industry shocks, we find that losses generally propagate via the liability and ownership channel and the reverse liquidity channel. But when the losses are large enough, the fire-sale effects dominate. For bank reputational shocks, we simulate a loss of confidence in major banks via deposit withdrawal and fund redemption. While the overall losses are much smaller than those of industry shocks, these risks cannot be ignored since the mutual fund industry stands to suffer and panic selling could amplify the losses.
Should All Blockchain-Based Digital Assets Be Classified Under the Same Asset Class?
The literature is well aware that blockchain-based digital assets would constitute a new asset class. However, it has been rather silent about the distinction among them. This paper discusses the digital tokens’ differences and similarities by their (i) creation and initial distribution; (ii) intended properties; (iii) actual usage; and (iv) behaviors. Although the digital tokens are indistinguishable in some aspects, they differ in the way they are created and initially distributed. Some of them have distinguishable risk and return profiles. Therefore, we take a view that the digital tokens take (or will take) different roles in the financial systems; should be classified under different asset classes; and should be subject to different sets of regulations (although some may overlap).
The Impact of Regional Isolationism: Disentangling Real and Financial Factors
Recently, there is a pressure for isolation policies both within the United States and among the EU members. The pressure arises due not only to the difference between regions in the U.S. and/or countries in the EU, but also to the difference across their population which affect the gains and losses from economic integration, both real as from trade in a common market and financial as in a monetary financial union. To get a better understanding of this pressure, one would need a model of trade and capital flows that takes into account the difference between individuals in a region and differences across regions. There is also a need for detail data at the individual and aggregated level, which often are not available. In this paper, we use unique long-panel data of households in Thailand, and from these data, we construct the household financial accounts, the village economic accounts, and the village balance of payments account. We also provide stylized facts on factor prices, factor intensities, financial obstacles, and village openness document differences across regions. Finally at the national level it is clear there is co-mingled variation in trade via devaluations and in finance via policies toward off shore bank and within-country financial infrastructure.
We develop a heterogeneous-agent/occupational-choices/trade model with financial frictions carefully built up and calibrated around micro and regional facts, that is, at both the individual level and the aggregate level. Then, we conduct two counterfactual policy experiments. In the first counterfactual experiment, we distinguish the effects of trade from the effects of capital flows. More specifically, we determine what would happen if we allow the prices of goods to change as in baseline scenario while keep borrowing limits and interest rates constant, and vice versa. In the second counterfactual experiment, we determine the effect of isolation policies that impede trade and/or capital flows across regions. We find through these counterfactual experiments that both real and financial factors are at play, that there are differences across regions in impact even when (policy) movements in variables such as interest rates and relative prices, which are exogenous to the regions, are common; impacts can be large, and vary with policy; and impacts are significant heterogeneous with both gains and losses and non-monotone movement across wealth classes and occupations, even allowing for occupation shifts which apriori might have mitigated impact.
Foreign Exchange Order Flows and the Thai Exchange Rate Dynamics
Applying the microstructure approach to exchange rates, this paper aims to shed light on the price formation process in the Thai foreign exchange market using a unique supervisory dataset of daily foreign exchange transactions from all licensed dealers in Thailand. We examine the main drivers of different types of order flows and the effect of resident and non-resident customer order flows on the Thai exchange rate. The results suggest that non-resident order flows have an important influence on movements in the Thai baht, while resident order flows do not. Regarding investors’ trading behavior, we find that non-resident order flows are driven by both fundamentals and movements of the Thai baht. Specifically, non-resident players appear to be ‘trend-followers’ with regard to exchange rate returns, exerting buying pressure when the baht recently appreciated. In contrast, domestic players tend to behave as ‘contrarians’, by buying the Thai baht after it depreciates.
The Journey to Less-Cash Society: Thailand’s Payment System at a Crossroads
Digital technology is changing the way we transact and pay each other, but cash usage remains dominant in many countries. In Thailand, it remains a question whether and to what extent electronic payments (e-payment) can replace cash. What is the role of a central bank amid challenges and opportunities at this crossroads? The paper explores global trends in cash and e-payment and outlines Thailand’s existing retail payment landscape. Both physical and IT/ICT infrastructure are assessed at micro-level with regard to Thailand’s readiness to move away from cash. However, given coexistence of cash and e-payment at present, we explore ways in which efficiency of cash management process can be improved. Data on cash distribution by geographical area are utilized to illustrate usage of Thai consumers and identify costs and inefficiency associated with cash management. On the other hand, adoption of e-payment can play a critical role in moving toward a less-cash society, if not a cashless one. The paper highlights the latest data on e-payment behavior in Thailand, especially PromptPay transactions as well as mobile/internet transactions after the transfer fee reduction in March 2018.
Institutional Capital Allocation and Equity Returns: Evidence from Thai Mutual Funds’ Holdings
Information about mutual funds’ stock holdings can provide useful signal for investors. In this study, we show that portfolio of stocks that are not favored by mutual funds tend to perform poorly, with monthly returns of 0.38% to 0.82% lower than stocks more widely held. When compared against asset pricing models, portfolio of such stocks can have monthly alphas as low as -0.33%, and the reason seems unrelated to stock-picking ability. One possible explanation is that demand from institutional investors can drive up stock prices, highlighting the importance of investor clientele in emerging market asset pricing.
Chasing Returns with High-Beta Stocks
One of the proposed explanations for the low-beta anomaly – a prevalent yet puzzling empirical finding that stocks with low systematic risk tend to earn higher returns than the Capital Asset Pricing Model (CAPM) predicts and vice versa – is that leveraged-constrained and index-benchmarked mutual funds drive up demand for high-beta stocks, leading to systematic mispricing. We find evidence that Thai mutual fund managers, on average, favor high-beta stocks and tend to alter their portfolio composition of high-beta stocks in response to fund flows. In addition, funds that hold high-beta stocks perform poorly compared to their peers: a one standard deviation increase in high-beta stock holdings is associated with a 1.3 percentage point decrease in future relative returns.
Value Investing with Quality in the US Public Insurance Companies
This study explores the value investing strategy coupling with quality metrics for the U.S. insurance industry. It uses apparent measures of insurance company efficiency such as loss ratio, expense ratio, combined ratio, and investment yield to construct portfolios. There are evidences of value premium as measured by PB and PE ratios. It is not clear that the quality metrics can give superior returns for investors. The anomalies can partially be explained by Fama-French five-factor model (FF5)’s market factor, value factor and profitability factor. The study also proposes using a new five-factor model that changes the profitability (quality) factor slightly from the Fama-French five-factor model. The adjusted FF5 “local” using insurance local factors do not improve the ability to explain the portfolios’ returns.