ABSTRACT
Trading algorithms require three essential ingredients: data, technology, and models. Of the models needed for algorithmic trading, price impact models have a unique, direct effect on the profitability of trading in the real world: price impact models establish the actual cost of trading at scale. Price impact analyzes trading's effect on stock prices. The study of price impact is distinct from the study of the stock's underlying value: economic theory models the long-term behavior of stock prices using rational expectations and fundamental efficiency. Economists have built so-called toy models of trade-induced price dislocations since the 1980s. The Econophysics community tested price impact models using high-frequency data and led a series of empirical discoveries and practical implications for trading. Trading data depends on the internal systems of financial institutions and the messaging format of trading venues. Trading strategies keep track of trading parameters to guide their decisions.
